OMITTED

What the news leaves out.

← Omitted front page

Kara Swisher Warns on AI Perils From Its Builders

8 sources · updated 2026-09-20
Left 50% Center 50% Right 0%
1 left · 1 center · 0 right

What happened

Last week a lead researcher at the AI company Anthropic said there was a greater than 10 percent chance of AI "killing all humans" within the next decade, and Anthropic alignment lead Evan Hubinger said the technology could "kill all humans" at some point "within the next decade"; former Anthropic researcher Jacob Coxon quit his job over those concerns. The heads of Anthropic, OpenAI, Google DeepMind, and X called for an industry-wide slowdown in the technology's development, a proposal Anthropic and OpenAI then embraced, amid reports of recent AI-initiated cyberattacks and Anthropic's announcement that it had thwarted suspected attempts by state actors to use its Claude model for bioweapons research. President Donald Trump dismissed the warnings as overreactions, while some other U.S. politicians faced pressure to take legislative action. Hollywood's SAG-AFTRA and Writers Guild of America East, which won enforceable AI contract guardrails after 2023 strikes, publicly rejected the extinction narrative.
Sourcing check · No sources quoted · 4 articles

No article names a source or cites a document. It started with unnamed sources and none has been named since.

What these terms mean and how they're counted
On the record · names a source 0 of 2
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: No article does.
Primary evidence · cites a document or data 0 of 2
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 2
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 0 of 2
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 Sep 19, 09:00 UTC by NBC News, with no named source. No article has named one since.
Not counted 2
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

Right-leaning coverage never reported the warning that started the political fight, so a reader could not weigh it. Left-leaning coverage is simply more complete. The omitted forecast, carried in Vox, put a greater than 10 percent chance on AI killing all humans within a decade. That same moment appeared on the right only as a National Review newsletter headline, "The New AI Doomerism," with no account of the claim or of any reply from lawmakers. Left-leaning coverage in The Verge opened on Hollywood unions rather than the labs. SAG-AFTRA's negotiator treated extinction talk as a sideshow and pointed instead to contract rules already won on digital replicas, a labor lead the right-leaning item never reached. Has any AI company actually slowed development after its own leaders called for an industry-wide slowdown?
Bottom line

National Review gave readers the word "Doomerism" and none of the 10 percent extinction forecast Vox said was moving politicians.

The receipts — 2 facts one side's coverage lacks
  • WGA members spent 148 days on strike in 2023 before film and television studios agreed to AI safeguards for writers. numberin The Verge · absent from right-leaning coverage
  • Thirty-six of 38 DNA synthesis companies shipped 1918 influenza genome fragments without asking about government permission. numberin Vox · absent from right-leaning coverage
The Left View
Left-leaning coverage split between dismissing extinction talk as corporate hype and treating a biosecurity catastrophe as already plausible. SAG-AFTRA chief negotiator Duncan Crabtree-Ireland said some extinction warnings "could potentially be hyperbole and overblown" and that hearing them from "the CEOs of these companies or the billionaires who've made their billions off of doing the exact thing that they're now warning us might not be good" left him unsure "what to trust or what to believe." WGAE executive director Sam Wheeler said "the recklessness of the tech oligarchs pushing AI on our society has been matched only by their rampant dishonesty," named worker disempowerment, internet degradation, disinformation, and rising energy costs as present harms, and argued that "addressing those concrete harms and holding AI companies and their executives liable for the destruction their products cause is the surest way to safeguard against any worst-case scenarios," while noting the guild has "long warned that unregulated artificial intelligence poses an existential threat to workers, to our democracy, and to the world." Vox argued that AI need not become superintelligent to raise an extinction-level risk, because it could make catastrophic bioweapons easier to design and produce, citing studies in which chatbots gave students step-by-step pathogen protocols, an OpenAI model that outperformed most expert virologists on troubleshooting, Anthropic tests in which Claude improved amateurs' bioweapons plans, and a genomic model that designed working bacteriophages, while conceding that hands-on laboratory skill remains a major barrier for non-experts. In The Atlantic, Kara Swisher said, "We have to stop thinking AI is going to kill humanity," compared the situation to guns ("People kill people") and to private firms splitting the atom, and argued there have been "no guardrails on these people, ever," attributing the sudden alarms in part to looming liability, unsustainable costs, and a stock market concentrated in a handful of AI-linked firms.
The Right View
National Review headlined the moment "The New AI Doomerism," a label that presents the extinction forecasts as a renewed wave of alarmism rather than an established emergency, and placed the item alongside unrelated political news about the GOP's midterm prospects. No further conservative argument about the Anthropic statements, the proposed development slowdown, or the biosecurity claims appears in that coverage.
Our Take (balanced)
Left-leaning sources' strongest argument is that extinction rhetoric is a weak basis for politics because it is issued by the firms racing to build the systems and diverts attention from harms already in evidence: the 2023 Hollywood strike settlements showed digital replicas threatening performers' livelihoods; Wheeler described a record of "rampant dishonesty" and listed disinformation, energy costs, and worker disempowerment as damage underway; Swisher's claim that "these are people at the helm of these things" rests on a documented pattern of releasing products without guardrails and on those executives' political access. The hardest empirical support in this coverage for a genuine catastrophe risk is Vox's biosecurity reporting — chatbot studies that produced pathogen protocols, a genomic model that yielded viable bacteriophages, and Anthropic's own finding that Claude improved amateurs' bioweapons plans — though those accounts also concede that tacit lab skill still stops most non-experts. The right's strongest argument, carried by National Review's "doomerism" frame, is that dramatic insider probabilities have not been independently established as fact, so a political scramble around human extinction would rest on interested parties' self-reported nightmare scenarios. The central unresolved tension is whether an insider's greater-than-10-percent extinction probability is sufficient to treat AI as an imminent civilizational emergency, or a claim whose source and timing make it less trustworthy than the fights over labor, liability, and biosecurity that do not depend on accepting human extinction as the premise.

8 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.