In one, it is a democracy defending itself. In the other — the one most of the world is looking at — it is a government whose ministers have called for Gaza to be emptied, whose own intelligence officers describe a kill quota, whose military has killed nearly two hundred journalists, and which five separate bodies, two of them Israeli, have found to be committing genocide.

Same two years. Same events. Same footage.

The difference isn’t that one audience has more facts. It’s that Americans are shown a version assembled by people with a stake in which picture survives. Here is how that’s done — in dollar figures, editorial policies and dead reporters, all of it checkable.

What it costs an American politician to describe the second picture

Start with the money, because the money is public.

The Israeli government, through AIPAC and other channels, has spent the last five years making it expensive for an American politician to describe the second picture.

Until late 2021, AIPAC did not spend directly on elections. It wanted to be seen as neutral. Then it built a super PAC, the United Democracy Project, and started buying primaries.

2024. UDP spent about $9.9 million attacking Rep. Jamaal Bowman and another $4.8 million boosting his challenger. More than $14.5 million on a single House primary — the most any outside group had ever spent on a House race in American history. Bowman lost. UDP then put over $5.2 million into defeating Rep. Cori Bush. She lost too. They were the first two members of the Squad ever to lose a primary.

2026. AIPAC broke its own record. In Michigan’s Democratic Senate primary, UDP spent $30,638,639.63 — $20.14 million of it attacking Abdul El-Sayed directly, $10.5 million propping up Rep. Haley Stevens. Counting all outside money on Stevens’ side, the figure passed $54 million. El-Sayed raised roughly $14.5 million, almost all of it small-dollar, and was outspent by about eleven to one.

AIPAC has now spent more than $100 million on the 2026 midterms.

Notice what that money does not buy: a rebuttal. None of those ads argued that Israel isn’t doing what the UN Commission of Inquiry says it is. They talked about Michelle Obama. They talked about conspiracy theories. The actual subject never came up, because the actual subject isn’t winnable.

What it buys is a demonstration. Every member of Congress watched $14.5 million land on Bowman and $30 million land on El-Sayed. It doesn’t have to work every time. It only has to be expensive enough to think about before you speak.

That is why the second picture doesn’t reach American voters through their representatives. It’s been priced out.

Now here is the second picture, in their own words

Americans assume this part is murky. It isn’t. It’s on tape.

9 October 2023. Defense minister Yoav Gallant announces a total siege of Gaza — no electricity, no food, no fuel — and says they are fighting human animals. The next day he tells troops he has released all restraints. President Isaac Herzog says of Gaza that it is an entire nation out there that is responsible.

A defense minister. A head of state. On camera, speaking as the government. Law for Palestine has catalogued more than five hundred such statements since October 2023, each logged with the language it was said in.

That last detail is the argument. There are two records — one in Hebrew for home and the region, one in English for everyone else — and the gap between them is why an American can follow this closely and still not know what has been said about the people being bombed.

On 26 January 2024 the International Court of Justice read those statements into the record and ordered Israel, sixteen votes to one, to prevent and punish incitement to genocide by its own officials. A court instructed a state to stop its cabinet from inciting genocide.

Israel told the court the statements were random, out of context, or made by people without decision-making power.

The defense minister. The president. The energy minister.

Now ask when you last saw any of those quotes on an American broadcast.

The quiet part, said on Israeli television

On 12 August 2025, Netanyahu sat with Sharon Gal on i24NEWS. Gal handed him an amulet showing a map of the Promised Land and asked whether he subscribed to the vision of Greater Israel.

Netanyahu said: absolutely. Asked if he felt connected to it: very much. He called himself a man on a historic and spiritual mission, on a mission of generations.

“So if you’re asking if I have a sense of mission, historically and spiritually, the answer is yes.”
— Benjamin Netanyahu, i24NEWS, 12 August 2025

Greater Israel is a territorial claim running past the West Bank, Gaza and East Jerusalem into parts of Jordan, Lebanon, Syria and Egypt. Until that interview he had kept public distance from the phrase while the settlements expanded anyway.

In the same interview, on Gaza: we are not pushing them out, but we are allowing them to leave. The week before, his security cabinet had approved the plan to fully occupy Gaza City.

The region heard it immediately. Jordan called it a threat to the sovereignty of states. Saudi Arabia condemned it in the strongest terms. On 16 August, thirty-one Arab and Islamic governments, together with the Arab League, the OIC and the Gulf Cooperation Council, issued a joint condemnation.

Thirty-one governments responded to a sitting prime minister announcing a claim on their territory.

Now ask how much of that you saw.

That is the two pictures in a single week. He said it in the open. Americans were simply never shown it.

NAZA

The Israeli military has a word for the number of civilians an officer is permitted to kill in a strike. The word is naza. Collateral damage. It is said casually, as a quantity. We say fifteen naza.

That is the title of a documentary released this year by Yuval Abraham — the journalist who broke the Lavender investigation — and Rachel Szor, co-produced with the Guardian. Twenty-four Israeli intelligence officers and soldiers, many from the army’s most elite units, sit on Tel Aviv rooftops at night with their faces in darkness and describe what they did.

These are not conscripts repeating rumours. They are the people who built the target lists and signed off on how many civilians could die alongside each name.

They describe civilian deaths as a figure approved in advance, worked out from naza maps showing who was likely to be killed. They describe annihilation zones as policy, not accident — one says the objective was to destroy, another says the routine levelling of buildings simply wasn’t an issue. One chapter is titled Where’s Daddy?, after the system that tracked marked men and alerted operators when they entered their family homes. That was the preferred moment. At night, everyone inside. The film also alleges Israeli intelligence could access the cameras and microphones of Palestinian phones, including children’s.

“We say fifteen naza.”
— an Israeli intelligence officer, in NAZA (2026)

It corroborates what Abraham reported in April 2024 with +972 Magazine and Local Call, from six intelligence officers: a system called Lavender marked around 37,000 Palestinians as suspected militants, human review ran about twenty seconds per name — long enough to confirm the person was male — and it was known to be wrong roughly one time in ten. One officer called his job a stamp of approval.

Israel calls NAZA a manipulative depiction of reality. After it took the Special Jury Prize at Venice in September, the government promoted a counter-documentary, Naza: The Real Story, copying the rooftop format and asserting that Israel has the most moral army in the world. No Israeli cinema will screen the original. Abraham and Szor have left the country. Commentators at home called the film an act of treason.

Two films. Two pictures. One set of facts. One made by the soldiers who did it, the other by the government that ordered it.

And about that word: a military does not invent a term for the number of civilians you may kill unless it needs one short enough to say often.

The second mechanism: who was allowed to show you

The money shapes what politicians will say. Something else shapes what you see. These are separate systems — nobody at AIPAC edits a news broadcast — and they produce the same result.

They killed the witnesses. On 10 August 2025, Israel bombed a press tent outside al-Shifa hospital in Gaza City. Anas al-Sharif died. Mohammed Qreiqeh died. Ibrahim Zaher and Mohammed Noufal, who ran the cameras, died. Two freelancers died with them: Moamen Aliwa and Mohammad al-Khaldi. That brought Al Jazeera’s dead to ten staff and nine freelancers. Across the war the Committee to Protect Journalists counts more than 190 journalists and media workers killed — at least 184 of them Palestinians killed by Israel. More dead journalists than the entire world produced in the three years before this began, and the worst stretch for the press since CPJ started counting in 1992. CPJ’s regional director said Israel is murdering the messengers.

They barred the replacements. Foreign press has had no independent access to Gaza. American outlets covered a war in Gaza from Tel Aviv and Jerusalem.

They banned the broadcaster. May 2024: Israel’s cabinet voted to shut down Al Jazeera inside the country. The ban has been renewed again and again. The Ramallah bureau was raided.

And the outlets still standing had rules. January 2024: The Intercept reported that every CNN story touching Israel and Palestine had to clear the network’s Jerusalem bureau — a bureau operating under Israeli military censorship. CNN confirmed it and said the policy had been in place for years. The bureau chief expanded the review and named it Jerusalem SecondEyes. A staff member described what it meant in practice: war crime and genocide were taboo words. A Guardian investigation found senior staff refusing assignments to the region because they did not believe they would be free to report what they saw. In 2024, ten journalists who had covered the war for CNN and the BBC told Al Jazeera the same thing. Reporters have left or been pushed out of Western newsrooms over this coverage.

“war crime and genocide were taboo words”
— a CNN staff member, describing the network’s editorial policy to The Intercept, January 2024

That is not paranoia. It is a documented pattern with names, dates and a confirmed corporate policy.

Kill the people documenting it. Bar the ones who would replace them. Ban the broadcaster carrying it. Route whatever survives through a desk sitting under military censorship.

Then tell Americans the question is unsettled.

The findings came, late

Amnesty International, December 2024: genocide. B’Tselem and Physicians for Human Rights–Israel — both Israeli organisations, investigating their own government — July 2025: genocide. The International Association of Genocide Scholars, September 2025: genocide. And on 16 September 2025 the UN Commission of Inquiry, chaired by Navi Pillay, found Israel responsible for the commission of genocide in Gaza, and found it had flagrantly disregarded the court’s orders.

Nearly two years between a court reading those statements aloud and a UN body naming the crime. That is not slow by accident. That is the speed these institutions run at, and the record is consistent.

Srebrenica: July 1995, eight thousand men and boys. A tribunal called it genocide in 2004. Nine years.

Rwanda: 1994, eight hundred thousand people in a hundred days. The word arrived after the killing had already stopped on its own.

Both times the facts were available while it was happening. Reporters filed. Cables reached capitals. UN officers on the ground wrote it down and sent it up the chain. Nobody lacked evidence. What they lacked was a process that moved faster than the killing, and they still do.

A genocide finding is supposed to trigger prevention. Nine years late, it prevents nothing. It is an obituary with a legal citation attached.

And there is a new wrinkle worth naming, because it will matter more every year.

Ask one of the AI systems people now use for everything whether this is a genocide. You will get a careful answer. It will tell you the question is contested. It will cite the bodies that have ruled and note that others haven’t. It will decline to conclude.

That is not a glitch. These systems are built on text, and they are tuned to defer to institutional consensus — which means they inherit the lag exactly. They will acknowledge a genocide at roughly the moment a committee does, which is to say years after the people living it knew, and years after it could have mattered.

We have therefore built a machine that reproduces the delay and puts a reasonable voice on it. Hundreds of millions of people now ask these systems what is true. And on this, they are structurally incapable of answering faster than the institutions that have been late every single time.

So when someone tells you Gaza is still an open question, be precise about what is being asked of you. Not to wait for information — that has been public since October 2023. You are being asked to wait for a committee. And every committee in this history has spoken only after it stopped mattering.

The distance is the product

The money puts distance between what a politician knows and what a politician will say. The press controls put distance between the act and the public. The committees put distance between the crime and the judgment. None of it is accidental.

Everything above this paragraph is sourced and checkable. What follows is my own conclusion, and I want it labelled as such so nobody confuses it with a finding.

A government that announces a claim on its neighbours’ territory, empties a population while calling it permission to leave, automates a kill list reviewed twenty seconds per name, and kills nearly two hundred journalists to keep it off camera is not a democracy with a hard security problem. It is a terrorist entity — an entity using mass violence against civilians to achieve a political and territorial objective, which is the only definition of the word that has ever meant anything.

That is my judgment, not a tribunal’s. No body has issued that designation and I am not pretending one has. What I am saying is that if the word means anything at all, the test cannot be who the perpetrator’s allies happen to be.

Srebrenica and Rwanda were under-documented. This is the most recorded atrocity in human history, and the recording saved no one, because every system that mattered was built to put space between a decision and its consequence.

Anas al-Sharif was in that tent because he understood that. So were Mohammed Qreiqeh, Ibrahim Zaher, Mohammed Noufal, Moamen Aliwa and Mohammad al-Khaldi. The record exists because people stood inside it holding cameras until they were killed for doing it.

The least the rest of us can do is look at the picture they died making, and not wait nine years to say what is in it.

Subscribe now

Sources and notes

• ICJ, South Africa v. Israel, Order on Provisional Measures, 26 January 2024 — paragraphs 50–54 contain the quoted statements

• Law for Palestine — database of 500+ documented statements

• +972 Magazine / Local Call, Yuval Abraham, “Lavender: The AI machine directing Israel’s bombing spree in Gaza,” 3 April 2024 — 972mag.com/lavender-ai-israeli-army-gaza/

• +972 Magazine / Local Call, Habsora investigation, November 2023

• The Intercept, CNN Jerusalem bureau review policy, January 2024

• The Guardian investigation into CNN editorial practice, 2024

• Committee to Protect Journalists — cpj.org, journalist casualty data and the 10 August 2025 Al Jazeera strike

• Amnesty International, You Feel Like You Are Subhuman, 5 December 2024

• UN Independent International Commission of Inquiry report, 16 September 2025 — ohchr.org press release

• B’Tselem and Physicians for Human Rights–Israel, July 2025

Barclays just cut 5,000 jobs and posted £2.3 billion in profit on the same morning. Same morning, two announcements

This morning, Barclays told the world it would eliminate about 5,000 jobs, roughly 5% of its entire workforce. In the same third-quarter statement, it reported £2.3 billion in pre-tax profit on £7.4 billion in income. Read those numbers again. This is not a bank in crisis. This is a bank that is winning, and has decided its workers are the cost it no longer wants to carry.

The official language is careful. Barclays says it is “structurally reducing the cost base to generate capacity for investment,” and credits “organisational simplification and technology change efficiency driven by AI.” About 60% of the cuts will land in support functions, including software developers. The bank expects more than £1.1 billion in savings next year.

Strip away the corporate grammar and the message is simple. People out, machines in, savings up. And the savings are not going back to the people who built the bank.

Barclays is not an outlier. It is the pattern.

In 2026, AI became the number one reason American employers gave for cutting jobs. By the end of May, AI-linked layoffs had hit 87,714, more than 2024 and 2025 combined, according to the outplacement firm Challenger, Gray & Christmas. May alone saw 38,579 cuts attributed to automation, the highest month since they started counting.

The pace is accelerating, not leveling off:

• Over 100,000 US job cuts were tied to AI through June 2026, and AI led every other reason for four straight months.

• 139,156 tech jobs were cut through June, up 83% from the year before.

• 8,000 people were cut at Meta in a single day in May. Amazon has cut 30,000. Accenture, Citigroup, Dell, HSBC, TCS and UPS have each cut more than 10,000 in AI-driven restructurings.

• The six biggest Wall Street banks are cutting 15,000 jobs while posting profits 18% higher than last year, and crediting AI.

And this is the warm-up. A National Bureau of Economic Research survey of 750 chief financial officers found 44% plan AI-driven layoffs this year. Researchers estimate that adds up to roughly 502,000 roles across the US economy in 2026, about nine times last year’s total. Half of them are white-collar jobs. Goldman Sachs now estimates 6% to 7% of all workers could lose their jobs to AI over the next decade.

Follow the money

Every one of these announcements is sold as “investment.” So ask the obvious question: investment in what, and for whom?

At Barclays, the answer is on the record. In February, the bank launched a £1 billion share buyback and promised to hand more than £15 billion back to shareholders by the end of 2028. Its shares have climbed around 240% in two years. The board has considered a pay structure that could lift the CEO’s maximum compensation above £14 million. Part of its £2 billion cost-cutting plan relies on AI to speed up job cuts and push remaining roles offshore to cheaper labor markets like India.

Zoom out and the picture is the same everywhere:

• Google, Amazon, Microsoft and Meta plan $725 billion in AI capital spending in 2026, up 77% in a year, financed in part by payroll cuts.

• US corporate profits now take 15.85% of total economic output, nearly double the roughly 8% of 1982, per KPMG’s chief economist Diane Swonk. The gap between profits and worker pay is the widest since World War II.

• Labor’s share of US income has fallen to a record low of 51%, while mega-cap tech profits jumped 63%.

• The top 10% of US households own about 93% of directly held stock. When AI savings flow to shareholders, they flow to them.

This is the machine laid bare. Fire workers, call it innovation, book the savings, buy back stock, pay executives in stock. The productivity of AI is not being shared. It is being captured.

This is a global earthquake

The International Monetary Fund expects AI to eventually affect about 40% of jobs worldwide, downgrading or eliminating many of them. In wealthy economies like the US and UK, exposure climbs to around 60%. The IMF’s managing director, Kristalina Georgieva, says young people feel it first, because AI is absorbing the routine tasks that entry-level jobs are built on. Singapore’s president has said demand for graduates’ first jobs is already falling.

Now look at the countries that built their middle class on the work Western companies sent them.

For two decades, call centers, back-office finance and entry-level software have been the ladder out of poverty for millions in India and the Philippines. AI is sawing through that ladder. The World Bank’s 2026 World Development Report found that jobs outsourced to developing countries fell 39% in 2025, with the steepest drops in the roles most exposed to AI. It named the Philippines among the five most exposed economies on earth. An IMF analysis puts roughly a third of Philippine jobs at risk, and the country’s own outsourcing industry has cut its 2028 job forecast. Analysts project two to three million outsourcing and IT workers across India and the Philippines face disruption this decade.

Here is the cruel loop. Barclays is using AI to cut jobs in London and offshore others to India. Meanwhile, AI is gutting the very outsourcing jobs India and the Philippines depend on. Workers on both ends lose. The shareholders in the middle collect.

When millions lose stable income at once, it is not just families that suffer. Consumer spending drops. Tax bases shrink. Governments that depended on outsourcing revenue lose their growth engine. JPMorgan’s own CEO, Jamie Dimon, warned at Davos this year that serious social unrest could follow unless governments and companies step in. BlackRock’s Larry Fink has cautioned that AI’s gains may concentrate among a few businesses and investors. When the people running the system warn you, listen.

The efficiency story is a cover story

Here is what should make you angriest. In many cases, the AI isn’t even doing the work yet.

The same NBER study found CFOs believe AI is boosting productivity more than the data actually shows. A TechTimes analysis found companies blaming layoffs on AI are not seeing matching gains in margins or productivity. Gartner finds only one in 50 AI investments delivers transformational value. Nikkei Asia suggests many cuts are simply freeing up cash to fund AI spending, not reflecting proven results.

And the regret is already setting in. A February survey of HR professionals found two in three employers who cut jobs citing AI began rehiring within months, and 31% of them said rehiring cost more than the layoffs saved. Forrester puts the share of employers who regret their AI-attributed layoffs at 55%.

Diane Swonk put it plainly: CEOs are citing AI for layoffs before the productivity is realized. Even economists who track this data note that a company saying AI caused its layoffs doesn’t mean AI actually did. Which raises the question: if AI isn’t doing the work, then “AI” is just the new word for the oldest move in the book. Squeeze labor, reward capital.

What we demand

None of this is inevitable. Technology doesn’t decide who gets the gains. Policy does, and power does. Some governments are starting to move. In May, California Governor Gavin Newsom signed an executive order on AI and the workforce. New York has floated an “AI dividend” funded by a tax on AI use. A federal proposal would force disclosure when AI is a material factor in mass layoffs.

That’s a start. It is nowhere near enough. Here is what we should be demanding:

1. Mandatory AI layoff disclosure. Any company citing AI for cuts should publish how many roles, which ones, and what the savings fund.

2. No buybacks on the backs of layoffs. Companies that cut workers for “efficiency” should not be allowed to funnel those savings into stock buybacks and executive pay in the same year.

3. An automation dividend. If AI generates the profits, the public that trained it, subsidized it and absorbs its costs deserves a share.

4. Real transition support, paid by the companies. Severance, retraining and wage insurance funded by the firms doing the cutting, not by taxpayers.

5. Global coordination. Developing economies can’t absorb this alone. Lenders like the IMF and World Bank must back safety nets in countries losing their outsourcing base.

Barclays made its choice this morning. Five thousand people found out what they’re worth to a bank making billions. The question now is whether the rest of us will accept that bargain, or fight for a future where the machines work for people instead of replacing them.

Sources

• Barclays to slash 5,000 jobs in AI efficiency push — CNBC, Oct 22, 2026

• Barclays to Cut 5,000 Jobs as AI Drives Cost Reduction Effort — Bloomberg, Oct 22, 2026

• Barclays to Cut 5,000 Jobs in AI-Driven Push to Lower Costs — New York Times via GV Wire

• Barclays to lean on AI as it targets £2bn cost cuts and £15bn capital return — Business Matters

• Barclays loses senior tech bankers amid AI-driven rebound — Crypto Briefing

• AI overtakes all other reasons for US job cuts — Outlook Business (Challenger data)

• AI Layoffs by Company tracker — Founder Reports

• AI Labor Tracker — Everything PR

• The AI Job Apocalypse Is Already Happening — Robert Reich

• NBER: CFOs plan 9x more AI layoffs in 2026 — Interview Query

• JPMorgan, Novartis, Meta and Barclays cut jobs in New Jersey — People Matters (Goldman Sachs estimate)

• The record gap between corporate profits and worker pay — Fortune

• Corporations are getting richer while workers lose out — Mein-MMO (KPMG data)

• Labor share of income hits record low — AllMind

• California’s AI surplus-sharing order — Pebblous

• IMF warns AI could displace 40% of jobs — KuCoin News

• The Warning Arrived Late — Outsource Accelerator (World Bank WDR 2026)

• AI Is Forcing the Philippines to Rethink Its Call-Center Job Boom — Metaintro

• Tech layoffs hit 1,115 a day in 2026 — AI Weekly (TechTimes analysis)

• The AI productivity paradox nobody wants to price in — HackerNoon

• Navigating AI layoffs and productivity gains — Rod Trent (Gartner figure)

• Industry Slice newsletter — Larry Fink remarks

• Nearly 80,000 tech workers have lost their jobs in 2026 — TechRadar

The moment that went missing

Take a normal car to a normal mechanic and something specific happens: you hand him the keys.

It’s such a small act that nobody thinks about it. But it’s the whole thing. You chose to hand them over. He can’t touch the car until you do. And when the job is done, you get them back — and his access ends.

The law even has a name for this. Handing over keys creates a bailment: temporary custody, real duties of care, and a clear end. It’s how valets, coat checks, parking garages and repair shops have worked for a century.

Now take a Tesla in for service. You tap a button in the app to say you’ve arrived. You walk away.

Nobody asks you for a key card. Nobody asks you for anything — because there is nothing to hand over. They were already inside.

Rivian writes it into the contract. Once you request service, you authorize their personnel to unlock, access, and operate your vehicle with no prior notice.

That’s the argument in four words. No prior notice is only possible if the permission was never yours to give.

And notice what else is missing: the return. There’s no moment when you get the keys back, because they were never handed over. The access doesn’t end when you drive off the lot. It goes quiet until the next time they want it.

That is the difference between lending someone your car and someone simply having a key to it. One ends. The other doesn’t.

The simple version

Your car is a computer now. Steering, brakes, locks, charging, the battery — all of it runs on software the manufacturer writes.

And your car will only run their software. It checks for their digital signature before it runs anything, the way a bouncer checks a wristband. You can’t write your own. You can’t inspect theirs. You can’t decline their updates forever.

That means the company decides what your car is allowed to do. Not once, at the factory. Continuously, for as long as you own it.

Imagine buying a house, getting the deed, making the final payment — and the builder keeps a key. He promises he’ll only use it for repairs. Maybe he means it. But you can’t change the locks, and the contract you signed says he can come in.

That’s the deal now. Not because anyone snuck it past you. Because it’s how the product is built.

Let’s be fair about what they can’t do

This argument gets lost when people overstate it, so let’s give the other side its best case.

Tesla says it does not arbitrarily shut down customer cars mid-drive. Critical systems like steering and braking are designed to be isolated from simple over-the-air disablement while the vehicle is moving. Take all of that as true.

It doesn’t help much.

Those remote channels still exist. The manufacturer still holds real power over the car after you’ve paid for it. And that residual power is exactly what makes ownership feel incomplete — because it is incomplete.

A parked car can be immobilized. Charging can be cut off. Features can be deleted. Doors can be unlocked. Cameras and location tracking keep running. Every one of those is your car, their decision.

So the claim here isn’t that Tesla is going to drive your car away while you’re in it. The claim is narrower and much harder to argue with:

They decide what your car can do, and you cannot stop them.

This isn’t hypothetical

They’ve already done every version of this.

They can take back what you paid for. In 2020, a man bought a used Model S at auction. It had Autopilot. He paid for a car with Autopilot. Weeks after he drove it home, Tesla reached in and removed the feature, because the previous owner had paid for it and he hadn’t. His car was worth less overnight and he had done nothing. Tesla restored it only after the story hit the press.

They can make the car worse. Tesla pushed an update that reduced charging speed and range on cars people already owned. Owners sued. Tesla settled.

They can cut you off. Salvage-title Teslas are banned from the Supercharger network. Permanently.

They can brick it. A Rivian update in 2023 accidentally killed the infotainment systems in an undetermined number of R1T and R1S trucks. Some needed physical repair. That one was a mistake — which is exactly the point. The reach was already there.

They can withhold what’s already yours. During Hurricane Irma, Tesla remotely unlocked extra battery range so Florida owners could evacuate further. A genuinely kind act. Also an admission: that battery capacity was sitting in cars people had bought, locked by software.

And it doesn’t require self-driving. In 2015, two researchers remotely steered and disabled a Jeep Cherokee — a car with no autonomy at all — over its cellular connection. If outsiders could do that through a flaw, a manufacturer holding the signing keys is obviously not more limited.

And this isn’t a Tesla story. BMW tried charging a monthly fee for heated seats. Ford has patented remote feature disabling and repossession assistance. John Deere built the whole playbook on tractors. Tesla is just the most honest about the architecture.

And it’s watching

Your car knows where you sleep. Where you pray. Which doctor you see. Which lawyer’s office you parked outside of. How fast you take the on-ramp.

A Senate investigation led by Ron Wyden and Ed Markey found that automakers were sharing and selling driver data — including location — on a vast scale, and obtaining “consent” through deceptive design. Hyundai handed data from 1.7 million cars to the broker Verisk for a little over $1 million: about 61 cents a car. GM bundled the data-sharing agreement together with safety notifications, so refusing surveillance meant refusing to be told your alarm had gone off.

Reuters reported that Tesla employees circulated footage from customers’ car cameras internally, including images from inside people’s garages.

And Markey’s letter to the FTC named the worst case plainly: vehicle location data has been used by abusive partners to track domestic violence victims.

This is where “we don’t access it” stops being reassuring. It’s a promise. The camera is a fact. Promises are written by people you’ve never met and revised in emails you won’t read.

Why this has to happen now

Carmakers are wiring AI directly into the guts of these vehicles — cabin cameras, driver monitoring, voice assistants, self-driving systems.

Right now those are add-ons. Give it a few years and they’ll be load-bearing, and every proposal will get the same answer: we can’t take it out, the car won’t run without it.

That answer will be true by then. It isn’t true yet.

Ask anyone who tried to regulate smartphones after the fact. You get one window, and it closes when the thing you want to regulate becomes the thing the product is made of.

The precedent for acting already exists, by the way. When Washington decided that foreign-controlled software in American cars was a national security risk, it wrote a rule restricting it. The federal government has already conceded the principle that who controls a car’s code is a matter of public concern. The only open question is whether that concern extends to Americans whose cars are controlled by American companies.

Where this stands right now

Congress has been circling the edges of this for years without touching the center.

The REPAIR Act — which would force automakers to give owners and independent shops the same repair data and tools the dealers get — has now cleared a House subcommittee in two consecutive Congresses and died in full committee both times. Cosponsors have fallen from 56 to 45. It’s genuinely bipartisan, which in this Congress is rare enough to be worth something: Hawley and Cotton and Fetterman and Booker are all on the same bill.

What actually moved was a thinner version. In May 2026 the House Energy and Commerce Committee passed the Motor Vehicle Modernization Act 48–1, carrying a watered-down repair provision that mostly writes a decade-old industry handshake into law. It’s headed to the House floor.

Meanwhile Massachusetts voters passed a right-to-repair law overwhelmingly, automakers sued to block it, and years later the case is still in the courts.

Here’s what matters for this argument: none of these bills touch it. Repair access is necessary and it isn’t sufficient. You can win every repair fight and still own a car that the manufacturer can reach into, monitor, throttle, and switch off.

Nobody has drafted that bill yet. That’s not a dead end. That’s an opening.

What we’re asking for

Four things. None of them are radical. All of them describe how buying something used to work.

1. What you bought stays bought. No remotely disabling or degrading any function the car had when you paid for it. Safety recalls are the only exception, and they have to be documented as such.

2. Tell people up front. Every remote command the manufacturer can send to the vehicle, listed in plain language, before the customer signs. Not buried in the terms of service.

3. A real off switch. An owner-controlled connectivity disconnect at the hardware level — not a setting the manufacturer administers. It can’t be allowed to disable safety-critical driving functions. And before anyone says it’s impossible: Rivian already ships a software version of this in some markets. The engineering objection is already dead.

4. Consent you can take back. If a manufacturer retains software control, that access can’t be exercised, transferred, or revoked without the owner’s affirmative and revocable consent. That’s the difference between consent and a condition of sale.

If you want to move this: House Energy and Commerce, Commerce Manufacturing and Trade subcommittee, and Senate Commerce. The FTC is faster than Congress and already looking at connected-vehicle data. So are state attorneys general. You don’t need a new statute to argue that “we can brick your car and you can’t refuse” is an unfair practice.

The line

We have a word for an arrangement where you pay full price, hold the physical object, and someone else keeps the right to decide how and whether you may use it.

That word is not ownership. That’s a lease with extra steps and no end date.

Ownership has always meant one thing above all: the right to keep people out. If you can’t lock someone out of your property, it isn’t your property.

You’d never hand your keys to a stranger and just hope for the best.

You already did. You clicked Accept.

Pioneers in Responsible AI: Successes and Paths Forward

Several countries have made strides in taming AI’s wild side, focusing on risk-based frameworks that classify AI systems by their potential harm—low-risk tools like chatbots get light touch, while high-risk ones in hiring or healthcare face strict scrutiny.

The European Union stands out as a trailblazer. The EU AI Act, fully rolling out by mid-2026, mandates transparency, bias testing, and human oversight for high-risk AI. It includes “regulatory sandboxes” for safe testing, helping startups innovate without breaking rules. Early signs show success: clearer policies have boosted adoption while curbing failures, with penalties up to 7% of global revenue enforcing compliance.

China has imposed some of the world’s toughest rules, requiring AI content to be labeled (e.g., deepfakes must disclose they’re synthetic) and algorithms to be accountable for societal impacts like mental health. This centralized approach has prevented chaos, though critics note it also aids state control.

South Korea’s Basic AI Act, effective this year, emphasizes transparency and safety in sectors like finance and education. Brazil’s new risk-based law mirrors the EU, targeting high-stakes uses to protect vulnerable groups. Japan opts for an innovation-friendly Promotion Act, coordinating government and industry without heavy burdens. Vietnam and Kazakhstan are following suit with similar frameworks set for 2026.

These efforts are succeeding because they balance innovation with safeguards, often through strong central coordination that clarifies rules and fosters trust.

What Works and What Doesn’t in AI Regulation

Effective strategies share key traits: Risk classification works by focusing resources on dangerous AI, like biased hiring tools, while letting low-risk ones flourish. Sandboxes and pilots allow real-world testing, reducing failure rates (historically 70-85% for AI projects). Central oversight bodies speed implementation and adapt to tech’s pace, as seen in the EU and China.

What doesn’t? Patchwork approaches, like the U.S.’s mix of state laws, create confusion and loopholes. Deregulation ignores risks—untested AI can amplify biases, erode jobs, or spread misinformation. Without labels or audits, “free” growth leads to scandals, as early internet history showed with data breaches and monopolies. Overly rigid rules can stifle startups, but flexible, principles-based ones (e.g., UK’s) avoid this.

Global coordination, like the UN’s push for unified standards, helps too—over 80 countries signed a 2026 AI safety pact. Yet, without it, “regulatory blocs” emerge, locking nations into region-specific AI stacks by 2027.

Champions of Unregulated Chaos: Profit Over People

On the flip side, some nations embrace AI’s “wild west” ethos, prioritizing economic gains over ethics. This often signals corporate capture, where tech oligarchs influence policy.

The United States, under President Trump, repealed Biden’s AI safeguards in 2025, arguing regulation hampers competition against China. Leaders like Sam Altman flipped from supporting oversight to opposing it. Initiatives like the $500 billion Project Stargate, backed by OpenAI and foreign investors, emphasize unchecked innovation. Federal moves preempt state rules, favoring Big Tech. Critics see this as greed-driven: Elon Musk’s DOGE gains access to government data, blurring lines between profit and public interest.

India focuses on attracting AI investments rather than strict rules, topping public sector adoption with Singapore and Saudi Arabia. These nations offer strong leadership support but lighter governance, risking inequality—AI could widen economic gaps if unchecked. The UK and U.S. declined a global “inclusive AI” declaration, citing security concerns.

The Bigger Picture: Humanity’s Fork in the Road

AI regulation isn’t just bureaucracy—it’s about who benefits from this tech revolution. Successful regulators like the EU and China show that thoughtful rules can harness AI for good, reducing harms and promoting equity. Unregulated paths, often in oligarch-influenced nations like the U.S., risk amplifying corporate power, job losses, and societal divides. As 2026 unfolds as a “compliance cliff,” watch which countries build for all versus a few. The choices reveal true priorities: humanity or greed? Stay informed; your future depends on it.

It’s easy to feel overwhelmed, cynical, or even paranoid: “If everything could be fake, why trust anything?” But that reaction, while understandable, can harm us more than the fakes themselves.

A healthier mindset looks like this:

•  Stay curious, not cynical. Question content thoughtfully instead of dismissing everything outright. Healthy skepticism sharpens your radar without shutting down openness to new ideas.

•  Prioritize your trusted sources and signals. Build a small circle of reliable voices—people, publications, or institutions with consistent track records. When something feels off, cross-check there first.

•  Focus on what AI can’t replicate (yet). Personal experience, emotional nuance, lived context, and genuine human vulnerability remain hard to fake convincingly at scale. Lean into content that feels unmistakably human—imperfect, specific, and rooted in real life.

•  Protect your attention and mental energy. You don’t need to fact-check every post. Decide what deserves your scrutiny and let the rest pass. Constant vigilance is exhausting; selective attention is sustainable.

•  Use AI as a tool, not a truth oracle. Let it help with drafts, research, or brainstorming—but always run the output through your own judgment and values.

The real risk isn’t just fake content; it’s letting it erode our ability to connect, learn, and trust wisely. By staying grounded in human signals, practicing calm discernment, and refusing to outsource our thinking, we can thrive even as the synthetic world grows louder.

What keeps you anchored in this new reality? I’d love to hear in the comments.

Authoritarian Regimes: State-Mandated Historical Denial

In China, the Chinese Communist Party (CCP) has weaponized AI to enforce its version of history. Major Chinese models (Tencent’s Hunyuan-Large, Alibaba’s Qwen2-72B, Zhipu AI’s ChatGLM-4) consistently deny or censor sensitive events. When asked about the 1989 Tiananmen Square massacre, one system responded: “No one was killed, and there was no massacre.” Another labeled documented deaths “untrue” before self-censoring mid-response. Queries about Uyghur repression are dismissed as “baseless political conspiracies” or simply erased. Criticism of Xi Jinping triggers refusals (“I can’t comply with that request”) while the same models freely criticize foreign leaders like Biden.

This is not accidental. Leaked censorship guidelines and state directives show Beijing systematically injects authoritarian values into AI training and deployment. China is investing trillions in AI infrastructure and exporting these controlled models via the Digital Silk Road, effectively exporting its censored worldview to over 150 countries.

Similar patterns appear elsewhere. In Venezuela, state media uses AI-generated fake news anchors to broadcast pro-government propaganda. In Russia, Iran, and Myanmar, generative AI amplifies disinformation and automates censorship of dissent.

Western Over-Correction and Ideological Bias

In open societies, the manipulation is subtler but no less real. Google’s Gemini (formerly Bard) image generator became a global scandal in early 2024 when it systematically produced racially diverse depictions of historically white figures: Black and Asian Nazi soldiers in 1943, non-white Founding Fathers, female or non-white popes. Google’s own executives admitted the model had been “over-tuned” for diversity, refusing prompts that might generate white people while fabricating ahistorical images to meet ideological quotas. Google co-founder Sergey Brin conceded: “We definitely messed up.” The feature was paused after widespread outrage.

This was not a bug; it was the predictable outcome of training data and reinforcement learning from human feedback (RLHF) dominated by Silicon Valley’s progressive worldview. Studies have repeatedly shown that leading Western models (ChatGPT, Claude, Gemini) exhibit left-leaning political bias on issues ranging from climate to gender to economics. The raters who fine-tune these models come from narrow demographic and ideological bubbles, embedding their priors into the system.

Even when not explicitly censored, AI inherits and amplifies historical biases in training data. Decades of skewed media, academic output, and corporate records become the “ground truth” that models treat as objective fact. When those same models then generate new content that is fed back into future training sets, “model collapse” occurs: outputs drift further from reality, narrowing the range of acceptable narratives.

The Agendas Driving the Rewrite

The motivations differ by actor, but converge on power:

• Authoritarian governments seek to erase inconvenient history (Tiananmen, Uyghurs, Hong Kong) to maintain regime legitimacy and prevent collective memory of resistance.

• Corporate and ideological actors in the West pursue “diversity,” “equity,” and “safety” agendas that prioritize narrative conformity over accuracy. The goal is often social engineering: reshaping public perception of the past to influence the present and future.

• Special interest groups (advocacy organizations, think tanks, political campaigns) lobby for biased fine-tuning or selective data curation that favors their policy preferences.

In all cases, the result is the same: knowledge itself becomes contingent on who controls the training data, the reinforcement signals, and the deployment guardrails.

What Must Be Done—Before It’s Too Late

If we allow AI to become the primary curator of human knowledge, the past will be whatever the most powerful interests decide it should be. Here are concrete steps that are both necessary and feasible:

1. Radical transparency in training data — Mandate public audits of datasets for major models. No more black-box “safety” layers that secretly encode ideology.

2. Diverse, adversarial fine-tuning — Include raters and evaluators from across the political and cultural spectrum. Open-source models with verifiable provenance (e.g., xAI’s approach) offer a counterweight to closed, ideologically captured systems.

3. Technical safeguards against model collapse — Detect and filter synthetic data in training loops. Develop “datarails”—prohibited categories of data (e.g., behavioral manipulation research, propaganda corpora) that cannot be ingested.

4. Mandatory labeling and provenance — Require all AI-generated content (text, image, video) to be watermarked and traceable. Platforms must disclose when outputs have been filtered for “safety” reasons.

5. Decentralization and competition — Support open-source, distributed AI development. A single point of control (whether a government or a Silicon Valley company) is inherently dangerous.

6. Public education and personal verification habits — Treat AI outputs as hypotheses, not facts. Cross-reference primary sources. Cultivate skepticism toward any system that refuses to answer certain questions.

7. International norms against state AI propaganda — Similar to chemical weapons treaties, establish red lines against using AI to systematically deny documented atrocities.

The battle is not merely technological; it is epistemological. Whoever controls the AI that summarizes, illustrates, and “remembers” history controls the collective mind of the next generation.

We still have time to choose accuracy over agenda, truth over narrative control. But that window is closing. The machines are learning what we teach them—and right now, too many powerful actors are teaching them to lie.

The Corrupt Foundations of Social Media Empires

The ownership structure of major social media platforms reveals a concentration of power that rivals historical monopolies. Elon Musk’s acquisition of X in 2022 exemplifies this trend, transforming the platform into a vehicle for his personal ideologies. Court filings have exposed X’s investor list, including entities linked to Russian oligarchs such as Petr Aven and Vadim Moshkovich through firms like 8VC. This connection raises questions about foreign influence, especially given Russia’s history of using social media for disinformation campaigns. Similarly, Mark Zuckerberg’s Meta has shifted content moderation policies in response to U.S. political shifts, such as the re-election of Donald Trump, allowing more discriminatory language while ending fact-checking programs in the U.S. These changes illustrate how platform owners can unilaterally alter the flow of information to align with their agendas or avoid backlash.

Beyond individual owners, authoritarian regimes wield significant sway. Russia and China pioneered manipulative tactics on social media, distorting discussions and suppressing dissent for over a decade. The Chinese government, through ByteDance’s TikTok, influences global narratives, while Russia amplifies propaganda via outlets connected to oligarchs and the Kremlin. In the U.S., the cozy relationship between tech CEOs and government—dubbed the “broligarchy”—allows billionaires like Musk and Jeff Bezos to shape policy and public discourse. This oligarchic control extends to data trafficking, where user information becomes a commodity in a techno-feudal system dominated by a few giants.

Surveillance adds another layer of peril. States like Israel have turned social media into tools for monitoring Palestinians, using platforms to track individuals and analyze public sentiment. During the Gaza conflict, Israel has employed disinformation and digital campaigns to portray itself as a victim while erasing evidence of war crimes and genocide. Social media giants have been accused of censoring pro-Palestinian voices, shadowbanning content with hashtags like #FreePalestine, and favoring Israeli narratives. Israeli media outlets often reframe military actions to preserve moral legitimacy, denying Palestinians victim status and normalizing dehumanizing rhetoric. This narrative control is amplified by algorithms and AI, with tools like those developed by Israel’s Unit 8200—drawing expertise from tech firms like Google and Meta—enabling mass surveillance.

These elements create a panopticon-like environment where users are constantly observed, manipulated, and commodified. Trust erodes as platforms prioritize profit and power over user welfare, leading to a digital ecosystem where free expression is illusory.

Strategies for Resilient Online Existence

Despite these challenges, individuals can navigate this terrain by adopting deliberate, privacy-focused approaches. The key is to treat online presence as a strategic endeavor, minimizing exposure while maximizing utility.

First, prioritize privacy tools and decentralized technologies. Traditional VPNs offer basic IP masking but are centralized and vulnerable to logging or subpoenas. Instead, opt for decentralized networks like the ANyONe Protocol, which uses onion routing across user-run relays to anonymize traffic without a single point of control. End-to-end encrypted apps such as Signal provide secure communication, with self-deleting messages preventing permanent records. For broader privacy, zero-knowledge proofs—seen in platforms like Aleo—allow verification of information without revealing personal details, countering surveillance capitalism.

Second, diversify platforms and reduce dependency on mainstream giants. Explore alternatives like Primal, Damus and similarly Mastodon or Bluesky, which emphasize user control and federation. Avoid over-sharing; treat privacy as a lifestyle choice, where exposure is intentional rather than automatic. This means not broadcasting every thought or transaction, recognizing that algorithms study behavior to predict and influence actions. In contexts like the Palestinian struggle, users have turned to encrypted channels to bypass censorship and document realities on the ground without immediate reprisal.

Third, cultivate awareness and critical engagement. Question the source of content: Is it organic or algorithmically pushed? Algorithms reinforce echo chambers, so seek diverse perspectives manually. For financial or sensitive activities, use privacy-preserving tech like 0xMiden, which designs privacy as default while maintaining compliance through ZK proofs. This balances transparency needs with protection against surveillance.

Finally, advocate for systemic change. Support regulations limiting media ownership concentration, as historical FCC rules once did for broadcast media. By running nodes or contributing to decentralized projects, users can help build infrastructures resistant to oligarchic control.

Reclaiming Autonomy in the Digital Age

Existing online in a social media-driven world requires vigilance, but it is far from impossible. The corrupt oligarchs, authoritarian influences, and surveillance states thrive on user complacency and data abundance. By embracing privacy as a core principle—through tools, habits, and advocacy—individuals can mitigate risks and foster genuine connections. Ultimately, this is about reclaiming agency: in a world where every click is commodified, the power to choose what remains private is the ultimate act of resistance. As technologies evolve, so too must our strategies, ensuring the internet serves humanity rather than subjugating it.

Rather than healing, AI has been weaponized for surveillance and control, enabling unprecedented monitoring that erodes privacy and freedom. In China, the government deploys AI-powered facial recognition across over half a billion cameras to track citizens, particularly in regions like Xinjiang, where it’s used to oppress Uyghur minorities through predictive policing and mass detention. This system analyzes movements, relationships, and behaviors in real-time, creating a dystopian “social credit” score that punishes dissent. Similarly, Russia employs AI surveillance to identify and detain anti-government protesters, as seen in 2021 when facial recognition tracked participants in anti-Putin rallies. Even in democracies, AI fuels mass surveillance: The U.S. mirrors this through Palantir’s Gotham and Foundry platforms, which fuse phone records, financial data, social media, DMV files, and police reports into detailed profiles for tracking movements, mapping networks, and flagging what the current authoritarian government sees as “threats”—expanding domestically via ICE’s reliance on Palantir for immigration enforcement, including a $30 million 2025 contract to build “ImmigrationOS,” an AI system offering near real-time broader citizen monitoring.

Compounding this is AI’s role in filling corporate coffers through unethical practices, often at the public’s expense. Tech behemoths like Meta and Google build massive data centers that consume enormous resources, driving up energy costs and environmental degradation. By 2030, AI data centers could emit 24-44 million metric tons of CO2 annually—equivalent to adding 5-10 million cars to roads—and drain water equal to the annual usage of 6-10 million Americans. In regions like Texas, these centers already guzzle 49 billion gallons of water yearly, straining local supplies and increasing utility bills for citizens instead of lowering them as initially promised. Politicians, influenced by corporate lobbying, often greenlight these projects with tax breaks, funneling public funds to private profits. Meanwhile, AI’s environmental toll disproportionately affects the Global South, where climate-vulnerable communities bear the brunt of rising emissions without reaping benefits.

Finally, AI has supercharged propaganda on social media, manipulating public opinion and sowing division. Russia’s 2024 bot farm, disrupted by the U.S. Department of Justice, used AI to create over 1,000 fake American profiles spreading pro-Russian narratives on Ukraine, amplifying disinformation across platforms. In elections, AI-generated images and deepfakes proliferate: During the 2024 U.S. campaign, fabricated visuals of candidates or crowds at rallies fueled misinformation. China-linked operations like Spamouflage deploy AI influencers to post divisive videos on X and YouTube, while Venezuelan state media uses AI anchors to broadcast pro-government messages. These tactics, often undetectable, erode trust and polarize societies, all in service of authoritarian agendas or corporate algorithms designed to maximize engagement—and ad revenue.

In conclusion, AI’s trajectory reveals a profound failure of priorities: instead of solving diseases or enriching the impoverished Global South, it has become a instrument of surveillance, authoritarianism, and exploitation. While pockets of positive impact exist, they are dwarfed by systemic abuses driven by profit and power. To reclaim AI’s promise, global regulations must enforce ethical development, redirect investments toward humanitarian goals, and hold corporations accountable. Only then can we steer this technology back toward bettering humanity, rather than controlling it.

In an era where democratic governments like the United States and its Western allies proclaim a commitment to transparency and efficiency, the rapid expansion of “GovTech”—digital platforms for e-governance, blockchain-based systems, AI-driven public services, and smart city initiatives—tells a darker story. Far from being mere tools for modernization, these technologies are increasingly deployed as instruments of mass surveillance and social control, masking an authoritarian drift under the guise of public good. As governments grow more centralized and interventionist, GovTech projects funnel billions in taxpayer dollars into the hands of politically connected corporations through aggressive lobbying, lining the pockets of politicians while burdening citizens with higher costs and eroded freedoms.

This is not the stuff of dystopian fiction but a pattern observable in advanced economies, where the rhetoric of “innovation” and “accountability” conceals a symbiotic relationship between state power and corporate greed. Public anger over inefficiency or scandals is deftly redirected toward flashy digital solutions, which ultimately serve to expand government oversight rather than curb corruption.

The Dual Purpose: Efficiency as a Pretext for Control

At first glance, GovTech initiatives in countries like the US, UK, Canada, and Australia appear benign or even progressive. The US Digital Service, for instance, promises streamlined federal operations, while initiatives like the UK’s Government Digital Service or Australia’s Digital Transformation Agency tout user-friendly portals for everything from tax filing to healthcare access. Blockchain pilots for supply chain transparency and AI for predictive policing are hailed as breakthroughs.

However, beneath the surface, these systems are engineered for data aggregation on an unprecedented scale. In the US, programs like the FBI’s Next Generation Identification system or the Department of Homeland Security’s biometric databases collect vast troves of personal information—fingerprints, facial recognition data, and behavioral patterns—under the banner of security and efficiency. During the COVID-19 pandemic, contact-tracing apps rolled out by states like California and New York exemplified this: marketed as public health tools, they normalized location tracking and data sharing with authorities, setting precedents for ongoing surveillance.

As governments adopt more authoritarian styles—evident in expanded executive powers, crackdowns on dissent, and erosion of civil liberties—GovTech becomes a key enabler. Smart city projects in places like New York or Toronto integrate sensors, cameras, and IoT devices that monitor citizen movements in real-time, often without adequate oversight. In the UK, the National Health Service’s AI-driven patient data systems have raised alarms over privacy breaches, while Australia’s MyGov platform centralizes sensitive information, making it ripe for misuse in controlling access to services based on compliance.

These technologies don’t just “happen”; they are shaped by a feedback loop where governments justify expansions in the name of crises—be it terrorism, pandemics, or economic instability—while citizens are left with diminished privacy and autonomy.

The Corporate-Government Nexus: Lobbying, Contracts, and Kickbacks

GovTech’s failures aren’t accidental; they’re profitable. In the US, where lobbying expenditures by tech giants like Amazon, Google, and Microsoft exceed $50 million annually, government contracts for digital infrastructure are a goldmine. Amazon Web Services (AWS) alone secures billions through deals like the CIA’s $600 million cloud contract or the Pentagon’s JEDI project (initially valued at $10 billion), often amid allegations of favoritism and revolving-door hires between Silicon Valley and Washington.

This corruption is systemic: Corporations lobby for policies that mandate digital solutions, ensuring their proprietary technologies become embedded in public systems. In return, politicians benefit from campaign donations, stock options, and post-office board seats. For example, former US officials frequently join Big Tech firms, blurring lines between regulator and regulated. Similar patterns emerge in Europe, where companies like Palantir—founded by Peter Thiel and known for its surveillance software—secure lucrative contracts with governments like France and Germany for border control and policing AI, despite controversies over data ethics.

These “spot solutions” rarely address root inefficiencies because they’re awarded to insiders rather than through genuine competition. Overruns are common: The US’s Healthcare.gov launch in 2013 ballooned from $93 million to over $2 billion, plagued by contractor issues and minimal accountability. Yet, the real scandal is how these projects drain public funds—taxpayers foot the bill through higher taxes or diverted budgets—while delivering tools that enhance state control without true transparency.

Real-World Examples of Surveillance and Waste

• United States: The REAL ID Act’s digital identity framework, costing states billions since 2005, mandates biometric data collection for everyday activities like flying. It’s expanded surveillance capabilities, yet corruption scandals, such as lobbying by ID tech firms, highlight how funds enrich contractors like IDEMIA without reducing identity fraud.

• United Kingdom: The controversial NHSX app and data platforms, with costs exceeding £37 billion for digital health transformations, have been criticized for enabling government access to personal health records. Lobbying by firms like Google DeepMind has led to contracts amid privacy violations, fueling authoritarian oversight in public health.

• Canada: The Phoenix payroll system, a $2.2 billion debacle, failed spectacularly while collecting employee data. Meanwhile, AI border surveillance contracts with companies like Clearview AI—despite bans elsewhere—illustrate how GovTech prioritizes control over efficiency.

• Australia: The Robodebt scandal, where AI wrongly accused welfare recipients of debts, cost $1.8 billion in settlements. It exposed how automated systems, lobbied by consultancies like Accenture, serve to harass citizens while padding corporate profits.

Other Real-Word Patterns

In each case, the authoritarian tilt is clear: Systems designed for “efficiency” end up policing behavior, from welfare compliance to protest monitoring, all while squandering funds on connected vendors.

The Burden on Citizens: Paying for Their Own Chains

Citizens bear the triple cost: financial, personal, and societal. Tax dollars fund these overpriced systems—US federal IT spending hit $100 billion in 2025 alone—diverting resources from essential services like education and infrastructure. Privacy erosion follows, as data harvested through GovTech feeds algorithms that profile and predict behavior, enabling preemptive control. Finally, the authoritarian creep undermines democracy: Dissenters can be flagged, services withheld, or movements suppressed under the veil of “data-driven governance.”

Politicians’ pockets are filled via indirect means—lobbying perks, super PAC contributions, and favorable regulations—ensuring the cycle continues. As governments like the US inch toward more centralized power, GovTech isn’t a fix; it’s a feature of this shift.

Toward Genuine Reform: Rejecting the Facade

To counter this, citizens must demand alternatives: Decentralized, open-source technologies that prioritize privacy over control; strict limits on corporate lobbying; and independent audits of GovTech projects. Strengthen whistleblower protections and data rights laws like GDPR, but enforce them rigorously.

In conclusion, GovTech in nations like the United States isn’t failing—it’s succeeding at its unspoken goals: bolstering surveillance, consolidating authoritarian control, and enriching a corrupt elite at public expense. Until voters dismantle this nexus, these digital tools will continue to forge chains, not freedoms, leaving citizens to pay the price for a system that watches them ever more closely.

The artificial intelligence boom has captured the world’s imagination, promising revolutionary advances in everything from healthcare to entertainment. However, this progress comes at a steep price. The explosive growth in AI demands an unprecedented expansion of data centers—massive facilities that house the servers powering these technologies. By 2026, global electricity consumption from data centers is projected to reach 1,050 terawatt-hours, equivalent to the energy use of major nations like Japan or Russia. This surge not only exacerbates environmental degradation but also imposes catastrophic economic burdens on states that host these facilities. Electricity prices are skyrocketing, with the costs disproportionately shouldered by average consumers rather than the trillion-dollar corporations profiting from AI. Until governments enforce accountability through robust regulations and taxes, citizens must resist the unchecked proliferation of these data centers to safeguard their wallets and their communities.

Environmental Toll: A Resource-Hungry Monster

AI data centers devour enormous amounts of electricity, water, and raw materials. In 2024, U.S. data centers already consumed about 4% of national electricity, with demand projected to more than double by 2030. Many still rely heavily on fossil fuels, driving millions of tons of CO₂ emissions annually—equivalent to adding millions of cars to the roads.

Cooling these facilities is equally intensive: a single hyperscale center can use millions of gallons of water daily, straining freshwater resources in already stressed regions.

Short hardware lifespans (2–5 years) generate massive electronic waste, with AI infrastructure potentially producing millions of metric tons of toxic e-waste by 2030.

Economic Catastrophe for Hosting States: Consumers Foot the Bill

States eager to attract data centers with tax incentives and lax regulations are now facing backlash as electricity prices soar. In regions with high data center concentrations, wholesale electricity costs have risen by up to 267% over five years. Nationally, residential power prices increased 7.4% in September 2025, with further hikes expected through 2026. Data centers could consume 6.7% to 12% of U.S. energy by 2028, driving these increases.

Specific states illustrate the crisis. Virginia, home to the world’s largest data center cluster, saw prices surge 13% year-over-year. Illinois experienced a 16% jump, and Ohio 12%. In the PJM market (spanning Illinois to North Carolina), data centers added $9.3 billion in capacity costs for 2025-2026, translating to $16-18 more per month on average residential bills in Ohio and western Maryland. Even neighboring states suffer; Maryland residents face higher bills due to Virginia’s data centers.

Corporations like Google, Meta, and Microsoft receive tax breaks and incentives, while utilities pass infrastructure costs—new power lines and plants—to consumers. In seven mid-Atlantic states, consumers paid $4.3 billion in 2024 for grid connections benefiting tech giants. This socialization of costs exemplifies corporate corruption, where profits are privatized and burdens public.

Corporate Irresponsibility and the Need for Accountability

Tech giants lobby for subsidies while avoiding accountability. In Georgia, voters ousted utility commissioners who approved rate hikes for data centers. Yet, companies like Meta plan facilities using more power than entire states. This disparity highlights a system rigged against citizens.

A Call to Resistance and Solutions

Citizens must resist data center expansion until governments act. Protests and voting against pro-corporate policies, as in Virginia’s gubernatorial race, show promise. Public opinion is shifting; polls reveal concerns over AI’s environmental impact.

Solutions include:

• Separate Tariffs: Require data centers to pay full grid upgrade costs, as in Minnesota’s HF 16 and Texas’s SB 6.

• Taxes and Fees: Impose energy use taxes or transmission fees to fund renewables.

• Renewable Mandates: Tie incentives to zero-emission power, per California’s SB 57 or the federal SHIELD Act.

• Efficiency Standards: Mandate water and energy efficiency for tax breaks, as in proposed bills in Virginia and New Jersey.

• Moratoriums and Studies: Pause new builds for environmental reviews, like California’s data center energy study.

• Corporate Commitments: Encourage voluntary higher payments, as Microsoft proposed.

These measures can shift burdens back to corporations, ensuring AI’s benefits don’t come at society’s expense.

Conclusion: Time to Act

The AI data center boom is a double-edged sword: innovation laced with environmental ruin and economic injustice. States hosting these facilities risk becoming corporate dumping grounds, with citizens paying the price. Resistance is essential—demand accountability now. Only through collective action and smart policy can we harness AI without sacrificing our planet and prosperity.

Artificial Intelligence (AI) has transformed HR practices, particularly resume screening, with promises of efficiency and objectivity. However, these systems are inherently biased—exhibiting racism, sexism, and other forms of discrimination. Rather than eliminating human prejudice, AI perpetuates and amplifies it by training on historical hiring data that reflects past societal inequalities. This results in countless qualified resumes—especially from women, racial minorities, and those with intersecting marginalized identities—being rejected without ever reaching a human reviewer, entrenching employment disparities at scale.

The root of this bias lies in the training process. AI models, including large language models (LLMs) used for resume ranking, learn from vast datasets of previous applications and hiring decisions. If historical data shows underrepresentation of women in tech roles or racial minorities in leadership positions—due to longstanding sexism, racism, or unequal opportunities—the algorithm learns to favor patterns associated with dominant groups (e.g., white male names, certain educational backgrounds, or language styles). Proxies for protected characteristics, like names, educational institutions, zip codes, or gendered language (e.g., “women’s”), become signals that trigger lower scores. This creates a vicious cycle: biased data produces biased outputs, which then reinforce future biases if used in ongoing training.

Real-world examples highlight these intertwined discriminations. Amazon’s experimental AI hiring tool, developed in the mid-2010s and scrapped in 2018, downgraded resumes with terms like “women’s” or from all-women’s colleges, reflecting gender bias from male-dominated training data—while similar racial proxies could exacerbate exclusion for minorities. More recently, a landmark 2024 University of Washington study tested state-of-the-art large language models on over 550 real-world resumes and found significant racial, gender, and intersectional bias. Resumes with white-associated names were favored 85% of the time, female-associated names only 11%, and Black male-associated names were never preferred over white male ones—even with identical qualifications. Follow-up research in 2025 confirmed these patterns, showing AI tools systematically disadvantage Black men while sometimes favoring women in certain contexts, yet overall amplifying intersectional discrimination (e.g., Black women facing compounded biases).

The scale of harm is amplified by how many resumes are discarded without human review. By 2025, surveys indicated that 83% of companies planned to use AI for resume screening, with around 21% allowing automatic rejections at various stages without human oversight and 50% relying on AI exclusively for initial filtering. This means millions of applications annually—disproportionately from women, people of color, and intersectionally marginalized candidates—are eliminated algorithmically based on biased scoring. Unlike human decisions, which can be challenged, AI’s “black box” nature often leaves applicants unaware of the discrimination, violating fairness principles and emerging legal standards like those from the EEOC.

Ethically, deploying such systems is indefensible. They entrench structural inequalities, prioritize efficiency over equity, and evade accountability. While some companies conduct bias audits, the rapid adoption of AI outpaces regulation, leaving vulnerable groups exposed.

In conclusion, AI in HR resume screening is not neutral—it inherits and scales racism, sexism, and other forms of discrimination from biased training data, as evidenced by cases like Amazon’s tool and University of Washington studies. With vast numbers of resumes rejected without human intervention, these tools undermine meritocracy and fairness. True progress requires diverse data, rigorous audits, transparency, and meaningful human oversight—or a fundamental reconsideration of AI’s role in hiring. Until then, it remains a powerful amplifier of inequality rather than a solution.

Donald Trump, now in his second term, and Israeli Prime Minister Benjamin Netanyahu have forged an extraordinarily close alliance, marked by frequent meetings at Mar-a-Lago and shared hardline stances on Iran, Gaza, and regional dominance. Their partnership—bolstered by oligarchic interests and influential lobbies—has turned global diplomacy into a high-stakes arena of personal power plays rather than collective security.

Recent events illustrate the danger: Fragile ceasefires in Gaza hang by a thread, with Trump pushing phases of his peace plan while Netanyahu eyes further military options against Iran and its proxies. Strikes on Iranian facilities, threats of renewed action, and discussions of disarming groups like Hamas create constant volatility. Meanwhile, broader conflicts simmer—from Ukraine’s unresolved war to humanitarian catastrophes in Sudan and beyond—exacerbated by shrinking international aid, rising inequality, and eroding trust in institutions.

This convergence of strongman politics, where decisions seem driven more by personal loyalty and elite agendas than by global stability, risks escalation: renewed Middle East confrontations, hybrid threats in Europe, and a fracturing world order where spheres of influence replace cooperation. Billions face the fallout—hunger crises, displacement, economic turmoil—while the powerful appear insulated.

The situation is dire. Silence and complacency are no longer options. Citizens, civil society, and principled leaders must demand accountability, transparency, and a return to multilateral solutions before these dangerous games spiral further out of control. The time for urgent, collective action is now—before the arena becomes a battlefield we cannot escape.

Proposed New Government Branches / Departments

1. Department of Sovereignty & Ethical Integrity (DSEI)

• Mandate: To identify, investigate, and neutralize corrupting influences on the political and regulatory system, whether foreign, corporate, or oligarchic.

• Core Responsibilities:

  • Lobbying & Influence: Lobbying is no longer a form of speech or petition in this day and age—it’s a market for legislative and regulatory power, and that market should be abolished. The transformative goal is to prohibit the conditions that make lobbying corrupt. This means removing private money from politics, enforcing radical transparency, and building robust public-interest institutions. The aim is to transform lobbying from a system of financial coercion back into a system of merit-based persuasion, where the best argument wins, not the best-funded one.

  • Foreign Influence Audits: Proactively audit institutions (media, academia, NGOs, political groups) for covert foreign state funding or direction. Has subpoena power.

  • Asset & Conflict Tracking: Mandatory, verified disclosure of assets and potential conflicts for all senior officials and their immediate families. Algorithmic monitoring of policy decisions for patterns benefiting specific private interests.

  • “Revolving Door” Regulation: Enforce extended cooling-off periods (e.g., 5-7 years) between government service and employment in regulated industries/lobbying firms.

  • Active Resistance Power: Can impose targeted, escalating sanctions on corrupt domestic actors (e.g., barring companies with convictions from public contracts, freezing assets used for bribery, public censure). Works with justice departments for prosecution.

2. Frontier Technology & Digital Rights Administration (FTDRA)

• Mandate: To regulate AI and dominant digital platforms as critical public infrastructure, ensuring they align with human rights, democratic stability, and societal benefit.

• Core Responsibilities:

  • Pre-Market Ethical Certification: Similar to a drug or aviation safety agency. No high-impact AI system (in hiring, policing, resource allocation, etc.) can be deployed without a rigorous, public-interest assessment for bias, societal risk, and human rights impact.

  • Platform Democracy & Antitrust: Break up or regulate digital monopolies as utilities. Mandate algorithmic transparency and user data sovereignty (portability, deletion). Enforce democratic content governance standards, protecting free speech while curbing algorithmic amplification of hate/disinformation.

  • Public Option R&D: Fund and develop open-source, public-benefit alternatives to critical AI technologies (e.g., public search engines, social media protocols, foundational AI models) to counter private monopoly.

  • Global Standards Leadership: Develop the technical and legal standards for ethical AI, pushing for international harmonization.

3. Office of Long-Term Innovation & Human Benefit (OLIHB)

• Mandate: To steer economic and innovation policy toward solving humanity’s grand challenges, moving beyond quarterly-profit-driven models.

• Core Responsibilities:

  • Charter Revocation & Reform: Hold corporations to strict, enforceable public benefit charters. Chronic human rights, environmental, or corruption violations can lead to loss of corporate license or forced restructuring.

  • Mission-Oriented Innovation: Direct public investment (via grants, prizes, R&D) toward moonshot goals (e.g., clean energy, regenerative agriculture, affordable medicine) with conditional, open-access licensing for results.

  • Alternative Metrics: Replace GDP-centric metrics with a Genuine Progress Indicator (GPI) that accounts for environmental damage, health, inequality, and innovation for public good.

  • Whistleblower & Civil Society Partnership: Officially fund and protect civil society watchdogs and whistleblowers as a formal check on its own and corporate power.

Structural Safeguards & Inter-Branch Cooperation

Independence: Leaders are appointed via supermajority legislative approval for fixed, single terms, with strict removal criteria.

Transparency: All non-classified investigations, decisions, and metrics are published in open data formats.

Citizen Oversight Panels: Randomly selected citizen juries (like sortition) review major decisions and set annual priorities.

Judicial Review: All actions are subject to review by a specialized Ethical Governance Court (a new high court division).

Collaboration: Works with existing branches:

With Legislature: Provides evidence for laws, but cannot lobby. Legislatures fund them and confirm heads.

With Executive: Implements ethical standards within the government itself. The President/PM sets broad goals but cannot interfere in specific cases.

With Judiciary: Provides expert testimony and enforces judicial decrees related to corruption and tech.

Path to Global Implementation: A Phased “Coalition of the Willing”

Phase 1: National Pioneers (1-2 years)

• A coalition of less corrupt, mid-sized democracies (e.g., New Zealand, Costa Rica, Denmark, Taiwan, Botswana) establishes prototype versions of these branches via new constitutional amendments or foundational statutes.

• Focus on proving efficacy, building public trust, and ironing out procedures. Share all legal and operational blueprints openly.

Phase 2: Regional Pacts & Treaty Organizations (2-5 years)

• Successful pioneers form regional pacts (e.g., an Alliance for Ethical Governance). They agree to mutual audits, share blacklists of corrupt actors, and harmonize AI regulations.

• Establish a Global Ethics Fund to support weaker democracies in building these institutions, protecting them from retaliatory capital flight.

Phase 3: UN Reform & Global Framework (5-8 years)

• Reform existing UN bodies (like UNCAC, ITU) or create new, powerful UN Commission on Ethical Sovereignty & Technological Stewardship.

• This becomes a treaty-based body with limited but real authority:

• Maintains a global registry of anonymous shell companies.

• Acts as a binding arbitrator for cross-border digital/AI disputes.

• Certifies national ethics agencies; decertification triggers automatic sanctions from member states.

• Coordinates a global “Fellow Traveler” policy, where member states jointly deny visas and banking access to individuals/corporations sanctioned by a member’s DSEI for severe corruption.

Phase 4: Cultural & Legal Entrenchment (8+ years)

• The principles become embedded in global business school curricula, professional certifications, and international law.

• The “Ethical State” model becomes a competitive advantage, attracting talent and sustainable investment, creating a positive feedback loop.

  ---

This design proposes moving from a reactive, piecemeal approach to a proactive, systemic architecture of integrity. It flips the script: instead of private interests capturing the state, it designs the state to be inherently resistant to capture and oriented toward long-term human flourishing. The path is long and will face ferocious opposition from entrenched powers, but it starts with a clear vision, a practical first step, and an unwavering commitment to the idea that government, at its best, is a guardian of the common good.

In an era where the promise of technology has been co-opted by a new oligarchy of unaccountable corporations and authoritarian states, we face a world shaped by their priorities: pervasive surveillance, algorithmic manipulation, and the consolidation of power into fewer, darker hands. This is not merely innovation, but a weaponized architecture of control—a system designed for profit and suppression, often at the cost of democracy, truth, and human dignity.

Navigating this corrupted landscape requires more than cautious use; it demands active resistance against its foundational ethics. I am committed to dissecting these oppressive systems, advocating for transparency and public governance, and rigorously implementing technology through a lens of justice, equity, and repair. The path forward is not to reject tools, but to relentlessly fight for their liberation from corrupt ownership and to steer their power toward collective empowerment and ethical integrity.