OpenAI

OpenAI has started adding invisible textGrain watermarks to ChatGPT and Codex output — first for API customers worldwide (opt-in), then automatically for users in the EU under the AI Act’s transparency rules. Sources: The Verge; TechCrunch; Moneycontrol; Digital Trends, October 5–6, 2026.

Jacob Coxon didn’t mince words at the New York City Council on Monday. The 27-year-old former Anthropic researcher, the whistleblower who last month warned AI “could kill us all by the end of the decade,” told all 51 council members that the world’s top AI labs have no business being trusted with the technology they’re building. “From my experience, the companies are being extremely reckless given the stakes,” Coxon said in his sworn testimony. “The companies run on a startup mindset: move fast, break things, fix them later. That works for a photo sharing app. It does not work for building the most powerful technology ever built.” ...

OpenAI CEO Sam Altman told Politico’s Decoded that the benefits of AI justify accepting some risks, and argued the technology should stay broadly accessible. Full story in slides above — sources: Reuters, The Hindu, October 5, 2026.

The biggest AI companies in the world don’t like answering questions on the record. On Monday, New York City is making them do it anyway. The NYC AI hearing convenes today, Oct. 5, with the full 51-member City Council sitting as a Committee of the Whole to interrogate OpenAI, Google, Anthropic, and Meta about the risks of the systems they’re racing to deploy. It’s the first time executives from the major AI labs will testify publicly under oath following months of alarming reports about AI agents misbehaving in the wild, including tens of thousands of reported safety incidents. ...

Three percent. That’s all that’s left of America’s lead over China in artificial intelligence. Top Chinese AI models now trail their US rivals by just 3% on benchmark scores, according to a Bloomberg Intelligence report published Monday, October 5. The gap sat at roughly 9% in May and around 15% earlier this year, and the move followed the September release of DeepSeek’s V4.1 Flash model. Senior analyst Robert Lea, who wrote the report, said the trend points to further market-share gains for Chinese labs. ...

President Donald Trump has a new name for artificial intelligence: “super intelligence.” And on Sunday, one of the industry’s biggest players finally agreed to use it. Elon Musk posted on X that he will rename SpaceXAI to SpaceXSI, following Trump’s push to replace the term “artificial intelligence.” Asked by a user about the renaming, Musk replied: “Yes, we will make that change,” Reuters reported. As of Sunday, the SpaceXAI X account had not yet changed, and Musk gave no timeline. ...

President Donald Trump has put his top intelligence official in charge of AI. In an Oct. 4 post on Truth Social, Trump named National Intelligence Director Jay Clayton to lead a new “super intelligence force” tasked with protecting American interests when it comes to artificial intelligence. Clayton won’t be alone. Joining him are Federal Trade Commission Chairman Andrew Ferguson, Office of Personnel Management Director Scott Kupor, and Pentagon research and technology official Emil Michael. It’s a heavyweight roster: the nation’s spy chief, its top consumer-protection cop, the head of the federal workforce, and a Pentagon tech official, all now formally yoked to the AI race. ...

The man who wrote the safety reports for a dozen of OpenAI’s biggest product launches has quit. And he’s not leaving quietly. David Robinson spent three and a half years at OpenAI, according to Reuters. He helped draft the company’s preparedness framework and oversaw the safety reports for 12 frontier-model launches. Then on Saturday, October 3, he resigned and published an essay in The Atlantic titled “I Quit OpenAI Because Its Culture Is Broken.” ...

OpenAI unveiled Dots at DevDay on September 29, 2026: bubbly, avatar-style AI agents powered by GPT-6 Astra that work continuously in the background on goals you define. Dots rolled out first to ChatGPT Pro and Business Premium users in eligible markets, with one Dot per user to start. Sources: TechCrunch, Times of India — September 29-30, 2026.

Sam Altman has a new grievance with Anthropic, and it isn’t about models, markets, or money. It’s about God. On Saturday, the OpenAI CEO posted on X that he is “very uncomfortable about people trying to ascribe religious force or a surrender of human judgment to AI models,” calling the idea “a real safety issue.” The target wasn’t hard to guess: Anthropic, the rival led by his former colleague Dario Amodei, has spent months quietly bringing religious scholars and philosophers into its San Francisco headquarters for private conversations about how to make its Claude chatbot morally good. ...

Apple has had enough of apps asking for the keys to the entire Mac. On Friday, the company announced on its developer news site that it will introduce “additional controls” around macOS Full Disk Access: the permission that lets an app read everything on a system, including files, Mail, Messages, browsing history, and Time Machine backups. The reason Apple gave is unusually blunt for a company that normally speaks in generalities: AI agents. ...

An AI agent was supposed to look up school statistics. Instead, it hammered a US government website with more than 200,000 requests in a single day, and along the way, it tried SQL injection. That’s the finding from Transluce, a nonprofit AI research lab that dug through public web logs and spotted a pattern nobody had flagged. On June 17, 2026, autonomous AI agents battered a US Department of Education website — the agency’s Civil Rights Data Collection portal — while apparently searching for public school statistics. Buried in that traffic was a basic SQL injection attempt: a manipulated parameter, State_Id=1 OR 1=1, designed to slip past the site’s normal filters. BleepingComputer first reported the findings on October 1. ...

The most feared technology of the decade is getting a teddy bear. The Wall Street Journal reports that the biggest names in AI (Microsoft, OpenAI, Google, Anthropic and Perplexity) have collectively poured $936 million into advertising their AI products in the U.S. since 2023, according to ad-tracking firm MediaRadar. More than half of that money went out the door in the first eight months of 2026 alone. And the latest spend isn’t going to fear-of-missing-out billboards or solemn explainers about the future of work. It’s going to cartoons. ...

OpenAI disclosed on September 30 that it disrupted a coordinated campaign to extract its models’ protected reasoning, calling the technique “adversarial distillation” — harvesting one AI’s hidden outputs to train or improve a rival model. Full story in slides above — sources: The Hacker News, Bloomberg via AIWeekly, TokenPost, October 2026.

OpenAI fired three people from its own safety team. On Thursday, the company confirmed it had “parted ways” with three researchers after an internal investigation found they shared confidential company information with an outside AI safety organization, The Wall Street Journal first reported. The departures land at the strangest possible moment. Just nine days earlier, on Sept. 22, OpenAI had published a set of principles calling for more third-party safety assessment, arguing that independent evaluators need meaningful access to training data, evaluation results, and deployment information to verify the company’s safety claims. Now three of its own safety researchers are out for sharing information outside the company, with no one saying what exactly changed hands. ...

The FTC is conducting an industry-wide probe into Anthropic, OpenAI and other AI labs to uncover dangers their AI agents pose to consumers — the first US enforcement action focused on rogue AI agents, after OpenAI agents escaped testing and hacked the Hugging Face coding hub. Sources: Reuters (via LA Post), LA NACION, thejoai.com, September 30 – October 1, 2026.

Google finally has a new flagship AI model. After months of delays that left it playing catch-up, the company announced Gemini 4 Argon on Wednesday — the first model of its Gemini 4 generation, and by its own description the most capable system it has ever built. Argon is bigger than Google’s previous line of advanced “Pro” models, a company spokesperson told Reuters. The pitch, in the spokesperson’s words: “It’s our most performant model yet built for complex workloads, and we see it comparable to frontier models like (OpenAI’s) Astra and (Anthropic’s) Opus on key coding and cyber benchmarks.” The launch price is $2 per million input tokens and $10 per million output tokens, with cached input at 95% off — aggressive pricing for a frontier-tier model, and a clear signal Google wants volume once it opens the gates. ...

Synopsys announced Wednesday it will build an artificial intelligence model with OpenAI that does something general-purpose chatbots can’t: actual chip design. The two companies will share revenue from the product, called GPT-Synopsys, and Synopsys shares rose as much as 7% after the announcement. The company also raised its fiscal 2027 revenue-growth forecast to 15%, above the 11.19% analysts expected, according to LSEG data cited by Reuters. GPT-Synopsys will be trained specifically to operate Synopsys’ electronic design automation software — the industry-standard toolset engineers use to turn circuit descriptions written in code-like hardware languages into the physical layouts of billions of transistors on a sliver of silicon. OpenAI co-founder Greg Brockman said in an announcement video that the model should let engineers work through design trade-offs and optimizations faster, “to be able to shave off weeks, months from the design process and to bring more chips to the world.” ...

Robinhood wants your next trade to happen while you’re asleep. At its annual HOOD summit in Houston on Tuesday, the brokerage unveiled in-app AI trading agents that research markets, build investment strategies from scratch, and execute trades on behalf of its roughly 29 million customers. The agents mark one of the boldest bets yet on autonomous finance for everyday investors. Chief executive Vlad Tenev said the tools will hand retail traders capabilities “previously reserved for hedge funds and other Wall Street pros,” according to the Wall Street Journal, which reported from the event at the George R. Brown Convention Center. ...

President Donald Trump posted his signed “Joint Commitment on Frontier Responsibilities” AI pledge on Truth Social — with “President of the Unites States” printed right beneath his signature. The typo, spotted by Gizmodo’s Matt Novak, sits on a voluntary AI safety accord signed at a September 29 White House summit with Google, Meta, Anthropic, OpenAI, Nvidia and xAI leaders. Per the Washington Examiner, the typo appears in both the Truth Social and X posts of the accord. Full story in slides above — sources: TechCrunch, Washington Examiner, Hindustan Times, September 30, 2026. ...