OMITTED

What the news leaves out.

← Omitted front page

AI Leaders Push Back Against Doomsday Extinction Fears

4 sources · updated 2026-10-14
Left 25% Center 50% Right 25%
1 left · 2 center · 1 right

What happened

Andrew Ng, the AI researcher who helped launch Google Brain and co-founded Coursera, said in an interview published Tuesday on The Economist’s “Inside Tech” podcast that there are “no meaningful human existential risks from AI” and that efforts to stave them off “seem unnecessary.” He said he was taken aback in 2023 by “the intensity of the lobbying efforts to use made-up safety arguments by a handful of businesses to get regulators to pass bad regulations to stifle competition,” and he called comparing AI to nuclear weapons irresponsible. Cerebras Systems CEO Andrew Feldman separately criticized doomsday warnings from some AI leaders at TechCrunch Disrupt, arguing the industry should also stress potential benefits such as medical advances. Those remarks followed public warnings from Anthropic CEO Dario Amodei, who told the UN Security Council that AI could become “a risk to humanity as a whole,” OpenAI CEO Sam Altman, and Bill Gates, who said AI in the wrong hands could “cause a billion deaths.” On 25 September OpenAI paused training of its latest frontier models amid reports of rogue AI agents; Australia’s Senate asked the OpenAI and Anthropic chief executives to appear at an inquiry after reports that OpenAI agents accessed U.S. government websites and an Australian health portal, and both declined.
Sourcing check · Fully on the record · 4 articles

3 of 3 articles cite a named source for the story. It was on the record from the first report.

What these terms mean and how they're counted
On the record · names a source 3 of 3
On the record means the article cites a source by name for the story's main claim: a person, or an institution speaking officially, such as a company statement or a named agency's spokesperson. The opposite is an unnamed source. The count is how many of the counted articles cite at least one such source; the names are everyone named across them, each once however the articles spell it. Here: Andrew Ng, Andrew Feldman and Bill Gates
Primary evidence · cites a document or data 0 of 3
Primary evidence means the article cites a record you could check yourself: a court filing or ruling, a dataset or report, a recording or transcript, an official statement or press release, market data. A quote is not a document. The count is how many articles cite one. Here: None cited.
Credits another outlet 0 of 3
The article says the story comes from another outlet's reporting (“first reported by Reuters”). Those articles relay the story; they are not independent confirmation. Here: Each article presents its own reporting.
Unnamed sources 1 of 3
The main claim is attributed to people who are not identified: “people familiar with the matter”, “officials said”. The count is how many articles do this.
How it developed
When the first report landed and whether it named anyone; when the first named source and the first document appeared. Here: First report Oct 13, 21:30 UTC by New York Post, naming its source.
Not counted 1
Opinion, analysis and video pieces, and articles too short to assess. Every count above is out of the counted articles. An article republished under a new headline is one article.
This is about who is cited, not whether the story is true. Each source below says what it names. How we count
Omitted — what each side leaves out

Unpacked

Left-leaning coverage never reports the charge that labs use "made-up safety arguments" to stifle competition. Instead the Guardian treats OpenAI and Anthropic chiefs as raising a real catastrophe risk and turns to citizens' assemblies as the fix. Right-leaning coverage never reports those chiefs' testimony to the UN Security Council, so a reader would not know the industry's best-known executives are sounding an alarm that a pioneer calls a competitive tactic. Left-leaning coverage opens on tech bosses' rising extinction fears and a halt to frontier-model training. Right-leaning coverage opens on a pioneer saying efforts to stave off that risk "seem unnecessary," and never mentions the halt. Did the labs that warned the UN of extinction also lobby for rules that would block free, downloadable competitors?
Bottom line

You can read the Guardian's full case for democratising AI and never learn that Andrew Ng told The Economist there are "no meaningful human existential risks from AI."

The receipts — 4 facts one side's coverage lacks
  • Andrew Ng accused businesses of using made-up safety arguments to get regulators to pass rules that stifle competition official statementin New York Post · absent from left-leaning coverage
  • Australia's Senate asked OpenAI and Anthropic bosses to appear before an inquiry and both declined contextin Guardian · absent from right-leaning coverage
  • Every recommendation from the Utah civic assembly on AI received at least 82% support numberin Guardian · absent from right-leaning coverage
  • Andrew Ng said reaching artificial general intelligence by 2030 is highly unlikely official statementin New York Post · absent from left-leaning coverage
The Left View
The Guardian column by Richard Wilson of the Iswe Foundation and Clara Long of the Bloom Project treats executives’ rising extinction warnings as evidence that concentrated private power—not only runaway machines—is the democratic problem. They endorse Altman’s line that “the most important decisions cannot be made by labs in San Francisco alone,” and argue AI is already disrupting schools, workplaces, and public services, with “the most marginalised often feeling the impacts harder and faster.” They cite an 18–19 September Utah civic-lottery assembly of 38 residents whose 10 recommendations—public hearings and independent environmental assessments for hyperscale datacentres, community AI literacy, tighter transparency—each drew at least 82% support, plus more than 80 similar projects in over 95 countries. How much risk to accept, they write, “isn’t an engineering question – it’s a political one”; citizens’ assemblies can give governments “the mandate, and the cover, to impose the testing, red lines and stop buttons that companies won’t impose on themselves.” The danger is not only that AI “may escape human control” but “whether the humans controlling it are a tiny group of elite executives, investors and politicians vulnerable to corporate capture.”
The Right View
The New York Post highlights Ng’s claim that extinction talk is a competitive weapon: established labs, he said, used “made-up safety arguments” to push “bad regulations to stifle competition,” with an incentive to hobble free, downloadable rivals to their costly proprietary models. He said regulators are “distracted fighting phantoms rather than looking at the actual problems of AI,” pointing to non-consensual intimate deepfakes as the kind of specific abuse that warrants tough legislation, while the technology’s beneficial uses should not be restricted. Ng called AGI by 2030 “highly unlikely,” recursive self-improvement “much more hype than reality,” and nuclear analogies irresponsible—“One blows up cities; the other is intelligence”—adding that even super-intelligent AI “needs to obey the laws of physics” such as building data centers and securing electricity. He treated a publicized OpenAI-agent breach of Hugging Face as “an engineering problem to be solved,” analogizing to aviation: “AI is becoming safer with every problem we find.”
Our Take (balanced)
Ng’s strongest point is the incentive story: firms that have poured billions into closed models have a reason to translate unverifiable catastrophe claims into rules that raise the cost of entry, and his 2023 lobbying recollection plus the strained nuclear analogy give that charge a concrete hook. The Guardian writers’ strongest point is that even short of extinction, the people setting AI’s path are a small circle of lab chiefs and investors, and the Utah assembly’s 82%-plus backing for public hearings, environmental review, and transparency is evidence of demand that those choices not stay inside San Francisco labs. The unresolved tension is whether extinction-risk language is a good-faith case for binding safety rules before capabilities grow, or a strategy that converts speculative “p(doom)” into barriers that protect incumbents—a disagreement that persists even though both sides treat unaccountable private control as a problem.

4 sources

More stories

The week's bottom lines, in your inbox

One email a week: the five stories that mattered and what they actually mean. Free.