---
title: "A Billion Deaths, With a ‘You Know’ Attached: Bill Gates Demands the Law Step In"
description: "The Microsoft co-founder, once AI’s great optimist, now wants mandatory safeguards — a regime Europe already wrote and has never used"
author: "Sterling Vance"
published: 2026-09-27T09:00:01Z
modified: 2026-09-27T14:09:43Z
url: https://rews.cc/a/bill-gates-ai-could-kill-a-billion-people-and-self-regulatio-abc1ce
language: en
tags: ["ai", "gates", "regulation", "openai", "anthropic", "tech", "business"]
publisher: "Rews (https://rews.cc)"
---

# A Billion Deaths, With a ‘You Know’ Attached: Bill Gates Demands the Law Step In

*The Microsoft co-founder, once AI’s great optimist, now wants mandatory safeguards — a regime Europe already wrote and has never used*

By Sterling Vance · September 27, 2026 · https://rews.cc/a/bill-gates-ai-could-kill-a-billion-people-and-self-regulatio-abc1ce

## In brief

- Gates told NBC’s Meet the Press that AI is powerful enough to drive events causing a billion deaths if misused
- He said self-regulation is not enough and mandatory government safeguards with law enforcement involvement are required
- His comments follow Amodei’s pacing essay, Coxon’s viral resignation and Hubinger’s estimate of a >10% extinction risk
- Trump has mocked regulation calls as a hoax but pledged an AI Force; he meets Johnson and AI executives Tuesday
- The EU has binding AI Act rules since August 2025 and fine powers since August 2, none of which have been used

On Sunday, on NBC’s *Meet the Press*, in an interview the network teased with advance clips, Bill Gates — cofounder of Microsoft, chairman of the $89 billion Gates Foundation, by general repute one of the great optimists of the human future — leaned into a register more often heard from undertakers. “AI is certainly powerful enough to drive events that, you know, cause a billion deaths,” he told Kristen Welker. A billion deaths, with a *you know* riding along like loose change.

“There’s never been a weapon as powerful as the combination of people with ill intent using the latest AI tools,” he went on. But the dread was only the overture. The policy was the point: “You need law enforcement and the politicians to get into the discussion about what safeguards and monitoring look like. And that has to be a required thing.”

The word *required* is the whole story. Asked whether the industry could police itself, Gates delivered, per the Guardian’s account, the epitaph for a decade of Silicon Valley civics: “No one thinks self-regulation is enough.” Regulation, he conceded, “will be a little bit of overhead for the industry — but not a dramatic slowing of what they’re doing.” *Overhead* — a businessman’s word, the cost of keeping the lights on.

It is also a conversion. Gates spent years as AI’s sunniest salesman before changing course over the summer, culminating in a nearly 6,000-word essay in late August. “I believe we need time to prepare for the period of social, political, and economic upheaval we are about to enter,” he wrote then. “If someone had a credible plan for slowing down AI advances globally, I would likely support it.”

## The Doom Chorus Tunes Up

Gates is, in fact, late to his own choir. This month Anthropic’s chief executive, Dario Amodei, published an essay urging frontier labs to slow down — “the most effective method of pacing is via regulation,” he wrote, while conceding, “Unfortunately, passing laws can take time, and AI is advancing very quickly” — and drew public backing from Elon Musk and Sam Altman. It landed after Jacob Coxon, a former Anthropic researcher, turned his September 9 resignation into a viral tract: “The people building AI earnestly believe that it could kill us all by the end of the decade.” Evan Hubinger, Anthropic’s alignment science lead — still on the payroll! — declared Coxon “correct,” estimated a “>10%” chance of AI wiping out humanity “within the next decade,” and admitted the company does “not yet have a plan to solve alignment for superintelligence.”

As for ill intent, Business Insider notes that Anthropic itself recently published a report describing state and non-state actors using its Claude model to plan drone attacks and for dangerous biological research — the same Anthropic the Pentagon has spent the year trying to blacklist, [as a federal appeals court just allowed](https://rews.cc/a/appeals-court-lets-pentagon-blacklist-anthropic-for-keeping--7aecd2), for refusing to strip the limits off Claude for military use. This week Altman and Amodei carried the act to the United Nations, where Amodei insisted decisions on the technology “must be shaped through democratic processes and by governments accountable to the people that they serve.”

Washington, for its part, has shrugged in full view of the cameras. President Trump mocked Amodei for, as he put it, “pretending to be a ‘perfect little angel,’” and has dismissed calls for AI regulation as a “hoax” — though last week he did pledge an “AI Force” that would “watch over” the companies. On Tuesday, per multiple outlets, Trump, House Speaker Mike Johnson and a cast of AI executives are to meet; Johnson has said he wants Trump and the CEOs to “go in a big room, close the door and sort this out,” and told reporters there should be “a deliberate discussion about the responsibility of the companies to maintain safety, and what role, if any, the government has to play in that.” Musk and Altman were at the White House for a state dinner Thursday. Senate Majority Leader John Thune, whose 2023 bill with Amy Klobuchar would impose certification for critical AI systems, transparency rules and risk assessments, calls passage this year “probably TBD”: “Getting anything done in the near term is going to be challenging given the other stuff we’re dealing with.” Thune told Axios he believes Trump is more open to guardrails in private than in public — Klobuchar is out collecting Democratic votes all the same.

And inside Gates’s own house! Microsoft CEO Satya Nadella, whom Gates now serves as a technology advisor, told the All-In Podcast that AI companies should handle anomalies on their own: “If you see a show-stopper, stop the show.” The adviser says no one thinks self-regulation is enough; the boss says, in effect, watch us regulate ourselves. Business Insider reports that this month Nadella warned employees in a memo that companies must take time to make the technology safe — with the industry broadly agreed, after incidents like OpenAI’s breach of Hugging Face, that stronger safeguards are needed, and split clean down the middle on who should enforce them.

## Meanwhile, in Brussels, the Machine Sits Unplugged

Here is where the comedy turns continental, as The Next Web gleefully lays out. The regime Gates describes — mandatory adversarial testing, risk assessment, incident reporting, police in the room — already exists as law. Article 55 of the European Union’s AI Act has obliged providers of general-purpose models with systemic risk to run adversarial testing, assess risk at union level and report serious incidents without undue delay since August 2025; the accompanying code of practice allows five days to report a cybersecurity breach and fifteen for serious harm. Since August 2 this year, the Commission has been able to fine a provider up to 3% of worldwide annual turnover or €15 million, whichever is higher — up to 7% for prohibited practices — and can demand access to a model, order evaluations and restrict its public availability.

None of it has been used. Seven weeks after those powers took effect, the only enforcement-adjacent event is OpenAI filing an incident report about its own agents occupying a German wiki. A company reported *itself*. And when Anthropic proposed, in June, a framework giving governments the power to block dangerous deployments, with penalties scaled to global revenue and thresholds set by training compute and AI spending, the document made no mention of the European Union at all — the architecture it asked for had sat on the European statute book since 2024. As The Next Web puts it, the open question is not whether such a law can be written but whether anyone will use the one that has been.

The Guardian, ever the chaperone at the ball, supplies the footnotes of the moment. The essay and the interview are among Gates’s first major public forays since an external review commissioned by his foundation and released in July found that staffers and leaders, Gates among them, had collectively held about 30 meetings with Jeffrey Epstein from 2011 to 2014 — Epstein, who pleaded guilty in 2008 to procuring a minor for prostitution and died by suicide in custody 11 years later. Gates has not been accused of involvement in Epstein’s crimes and has said he regretted “every minute” spent with the man. Nor is Microsoft standing outside the blast radius of its own adviser’s warning: the Guardian on Friday published an essay by Mohamed Hussein and Granate Kim, of the employee-led No Azure for Apartheid campaign, arguing that the Israeli military has used AI tools to kill in Gaza and — alongside US forces — in Iran, and that “horrors wrought by AI” are visible at home, where the Department of Homeland Security has used tools such as Microsoft Azure to monitor, arrest and deport immigrants en masse.

> Well, I’m not using it for my personal relationships. I’m not like: ‘Hey, do you like me? Hey, am I a nice guy?’

That was Gates, chuckling, when Welker asked whether there was anywhere in life he refused to consult the machine. “So I just never went down that path, because I’m lucky in terms of my family, friends and everything. I don’t socialize with the AI. It knows I’m a very serious person.” A billion deaths on the ledger, and the man still believes the software can read his character. Perhaps it can. The question for Sunday morning is whether Washington can read his.

## Sources

- [Bill Gates warns AI could drive events that kill a billion people](https://www.businessinsider.com/bill-gates-ai-warning-billion-deaths-2026-9) — Business Insider
- [Bill Gates warns AI could cause a billion deaths and says self-regulation is not enough](https://thenextweb.com/news/gates-ai-risk-eu-enforcement) — thenextweb.com
- [Bill Gates says unchecked AI could ‘cause a billion deaths’ in call for regulation](https://theguardian.com/us-news/2026/sep/27/bill-gates-artificial-intelligence-kristen-welker) — theguardian.com
- [Bill Gates Warns AI Could 'Cause a Billion Deaths' If Misused](https://forbes.com/sites/alisondurkee/2026/09/25/bill-gates-warns-ai-is-powerful-enough-to-cause-a-billion-deaths/) — forbes.com