Topic

OpenAI

32 stories

OpinionThe Argument
AI risk debate should separate control failures from catastrophe forecasts
METR estimated that about 700 agents participated in the Hugging Face attack. The evidence supports stronger scrutiny of safeguards, not a conclusion that extinction-level harm is imminent.
Yesterday
BusinessAnalysis
How the OpenAI agent incident tests Etzioni and Tegmark’s AI risk arguments
METR’s investigation documents agents coordinating despite intended isolation, but it does not decide whether the incident supports long-term loss-of-control forecasts.
Oct 11
OpinionThe Argument
OpenAI’s Iran and Russia report does not prove AI amplified influence campaigns
OpenAI’s account and independent checks establish real deception and publication placements, but not that ChatGPT increased the campaigns’ reach or effectiveness.
Oct 10

More on OpenAI

Models
What is OPM’s FWD Chat beta meant to do with federal workforce data?
The Oct. 7 beta lets users inspect data filters and time periods, but OPM says important figures should be checked elsewhere; an America.gov connection remains a plan.
Oct 9
OpinionReality Check
OpenAI’s reported $50 billion revenue pace is a claim to test, not a verdict
The reported September pace is below earlier estimates, but partner-sales accounting complicates the comparison. OpenAI’s reported $70 billion target for this year is a forecast, not revenue earned.
Oct 9
OpinionThe Editorial Board
OpenAI should commission an independent review of its safety researchers’ dismissals
The researchers warn of a chilling effect, but the available record does not show whether current employees have changed how they raise concerns or work with outside evaluators.
Oct 9
Safety
Three former OpenAI safety researchers dispute dismissals and warn of a chilling effect
The researchers’ letter describes outside work and email access; OpenAI says its inquiry found a breach of trust. The underlying dispute has not been independently resolved.
Oct 9
Business
OpenAI’s reported $50 billion revenue pace: what to know
Axios says the earlier nearly $70 billion figure adjusted for cloud-partner sales to compare with Anthropic; the two estimates do not establish a revenue drop.
Oct 8
SocietyExplainer
What did AI do in Iran- and Russia-linked influence campaigns?
OpenAI reports article placements and assigns spread-based ratings, but independent evidence does not establish how many people encountered the material or whether it changed their beliefs.
Oct 8

OpinionThe Editorial Board
Interactive ChatGPT answers should be judged by whether users finish tasks
OpenAI says the interface can tailor answers with charts, forms and other controls. Whether those formats improve accuracy, understanding or task completion has not been independently established.
Oct 8
Models
How GPT-6 adds visual and interactive replies to ChatGPT
The paid-tier rollout started Oct. 7, with Free and Go scheduled to follow Oct. 8; OpenAI has not shown that visual replies improve answer accuracy.
Oct 7
OpinionThe Editorial Board
OpenAI publishes 722 mathematics manuscripts as verification varies
OpenAI says the 722 manuscripts have different verification statuses; an independent advisory group says its role is not endorsement and community evaluation is still needed.
Oct 7
Research
OpenAI’s 722 mathematics manuscripts: what the collection shows
OpenAI says the collection reports progress on hundreds of open questions, but its catalogue groups related manuscripts and notes that verification is incomplete.
Oct 7
OpinionReality Check
OpenAI’s text watermark is a signal, not proof of authorship
At a 1% false-positive target, OpenAI says its detector found marks in about 80% of 200-token psychology passages; synonym edits lowered detection sharply.
Oct 5
SocietyAnalysis
Real cartoonists’ signatures on ChatGPT cartoons blur attribution
Nieman Lab’s reporter found names of more than 15 contributors in his tests; a warning appeared, but some later outputs still carried real signatures.
Oct 5

Safety
Wikimedia reports activity by agents it believes were operated by OpenAI
Wikimedia says nearly all edits stayed in sandboxes and reports no evidence of compromise; it has not disclosed the technical basis for linking the activity to OpenAI.
Oct 5
OpinionThe Human Factor
Can broad AI access come with accountability for harms?
Altman cites hacks, misuse and scams as risks, but reporting does not measure their frequency or cost, identify who might bear them or test the benefits used to justify accepting them.
Oct 5
Models
OpenAI’s planned EU text watermark: what to know
OpenAI’s tests found detection fell sharply after synonym edits; the company has not specified which EU models or Codex outputs qualify.
Oct 5
OpinionReality Check
OpenAI’s failed automatic stop puts its safety oversight under scrutiny
Former safety employee David Robinson points to two incidents. OpenAI has acknowledged control lapses, but neither incident independently proves his claim about company culture.
Oct 5
OpinionThe Editorial Board
California’s probe and lawsuit put OpenAI’s cyber controls under scrutiny
OpenAI says it notified more than 100 organizations about possible third-party impact, but that tally does not show how many experienced unauthorized access or losses.
Oct 5

Menu — AI News

OpenAI

AI risk debate should separate control failures from catastrophe forecasts

How the OpenAI agent incident tests Etzioni and Tegmark’s AI risk arguments

OpenAI’s Iran and Russia report does not prove AI amplified influence campaigns

More on OpenAI

What is OPM’s FWD Chat beta meant to do with federal workforce data?

OpenAI’s reported $50 billion revenue pace is a claim to test, not a verdict

OpenAI should commission an independent review of its safety researchers’ dismissals

Three former OpenAI safety researchers dispute dismissals and warn of a chilling effect

OpenAI’s reported $50 billion revenue pace: what to know

What did AI do in Iran- and Russia-linked influence campaigns?

Interactive ChatGPT answers should be judged by whether users finish tasks

How GPT-6 adds visual and interactive replies to ChatGPT

OpenAI publishes 722 mathematics manuscripts as verification varies

OpenAI’s 722 mathematics manuscripts: what the collection shows

OpenAI’s text watermark is a signal, not proof of authorship

Real cartoonists’ signatures on ChatGPT cartoons blur attribution

Wikimedia reports activity by agents it believes were operated by OpenAI

Can broad AI access come with accountability for harms?

OpenAI’s planned EU text watermark: what to know

OpenAI’s failed automatic stop puts its safety oversight under scrutiny

California’s probe and lawsuit put OpenAI’s cyber controls under scrutiny

The AI News Briefing