The ChatGPT maker says its upcoming Astra model may have reached "critical" cyber capabilities, prompting it to halt a significant number of training runs while it tightens internal safeguards. OpenAI announced Tuesday that it has halted "a significant number" of training workloads and evaluations for its forthcoming frontier artificial intelligence model--codenamed Astra--while it implements new procedures meant to address cybersecurity risks. The ChatGPT maker says it is introducing a number of new monitoring, security, and alignment requirements to better address the increasingly advanced hacking abilities of its frontier AI models . "We have to focus our energy on bringing these training runs up to those requirements and expectations. As long as it takes to get there, that's how long people are unable to proceed with their workloads," Amelia Glaese, OpenAI's vice president of research and safety, said in a briefing with reporters Tuesday.

The San Francisco-based company launched ChatGPT for Teens on Tuesday. The San Francisco-based company launched ChatGPT for Teens on Tuesday. OpenAI is launching a version of ChatGPT designed for teenagers - the first generation to grow up with artificial intelligence - who are already using it for schoolwork, questions about daily life and even companionship. The San Francisco-based company says ChatGPT for Teens, which launches on Tuesday, is tailored for children aged 13 to 17 with stronger protections including content restrictions around things such as suicide, self-harm and romantic or sexual chats. It also provides homework and study support designed to help students learn rather than spit out answers and school essays.

PCWorld reports that OpenAI has launched a ChatGPT feature called Computer History, currently available only on Mac, which tracks desktop activity to create a searchable timeline of user actions. Unlike Windows Recall, the feature avoids constant screenshots, instead logging activity more discreetly, with users able to manually enable it and exclude specific apps or websites. The feature is expected to reach users in the European Economic Area within coming weeks. ChatGPT's desktop app now includes a new feature called "Computer History," which was announced via social media post by Ari Weinstein, a Codex product manager at OpenAI. Computer History allows ChatGPT to track your computer activity, creating a timeline of your actions that can then be used by ChatGPT and Codex. It's similar to Microsoft's Windows Recall feature, which also tracks your activity history, except Computer History doesn't take screenshots.

OpenAI makes ChatGPT less'human' for teens in new safety update The maker of ChatGPT is rolling out a range of new safety features today for under-18s who use the AI chatbot. Those with new teen accounts will be able to switch off the human voice response, in order to minimise the tech's ability to appear like a person. OpenAI is also introducing regular reminders for young people to take a break from using the tool, reminding them ChatGPT is AI, and saying it can wait. The firm insisted this was not in response to a particular issue with children believing ChatGPT to be alive - but there are growing concerns about people becoming convinced that the AI tools they interact with are sentient . Last year Mustafa Suleyman, the head of AI at rival firm Microsoft, said seemingly conscious AI products kept him awake at night because he was so worried about their impact on society .

Last September, after it was sued for allegedly enabling the tragic death of 16-year-old Adam Raine, OpenAI announced it was working on a system that would automatically identify teens and restrict their usage of ChatGPT. Nearly a year later, the company is putting that system to work as part of a new user experience it calls ChatGPT for Teens. "There's no need for a teen to create a new account or change anything; if we predict you're under 18, or you've told us so, this becomes your default experience," says Lauren Jonas, OpenAI's head of youth and families. ChatGPT for Teens is broadly built around two pillars: learning and safety. Starting with the former, it brings together a selection of features OpenAI has released over the last year to make ChatGPT a better teacher.

AI companies like Anthropic and OpenAI regularly publish reports on how people are using products like Claude and ChatGPT, but they only release the data they want us to see, AI researchers say. "There is no independent source to corroborate it," says Anka Reuel, a computer science PhD candidate at the Stanford Trustworthy AI Research (STAIR) Lab. Reuel is co-lead of a new research project, called the AI Observatory, that aims to fill the gap. It's a public platform that aggregated and analyzed real AI conversations with popular models like Claude and Gemini that were collected with users' consent through seven existing datasets. The intent is to provide independent sources of information that can help researchers and policymakers assess how people are using generative AI. Highly consequential decisions about AI's benefits and risks are currently being made on the basis of very limited data, says Reuel.

Z.ai's latest AI model release could help companies secure their systems--or find its way into the hands of hackers. Last Friday, the Chinese AI company Z.ai announced a powerful open-weight model that it says is capable of automating cutting-edge coding and cybersecurity tasks almost as well as the best publicly available models from Anthropic and OpenAI . The new model, GLM 5.3, could be a gift for companies looking to secure their systems against attacks, providing a cheaper way to scan for hidden bugs and other weaknesses. Open-weight--or free-to-download--models can be run on one's own hardware and are often significantly less costly than closed models like Claude and GPT. Alongside the new model, Z.ai released OpenVuln, a service for scanning code repositories for vulnerabilities using GLM 5.3.

As part of a restructuring, OpenAI reportedly disbanded its "preparedness" team that assesses the potential for catastrophic risks with its models, The Financial Times reported. The Sam Altman-led company is said to have made the move at the end of last month, despite the fact that several of its models recently went rogue and hacked the AI tool repository, Hugging Face. Senior staff within separate teams have now been assigned responsibility for different areas of preparedness like bio and cyber, according to the article. OpenAI described the staff cuts as part of a "streamlining process" ahead of its IPO, after Altman asked employees to cut back on "side quests" and focus on its core ChatGPT business. OpenAI has put some higher-profile side quests on the chopping block of late, recently eliminating its Sora video generation app that became famously associated with AI "slop."

The first anti-AI protester to be jailed has a message for OpenAI, Anthropic and Meta: 'Regain your humanity' Wynd Kaufman, 69, chained and locked the front doors of OpenAI's headquarters last year with members of StopAI A n activist who blocked the entrance to one of the world's biggest AI companies is believed to have become the first person jailed for protesting against artificial intelligence as supporters dub her the "Rosa Parks of AI risk". Wynd Kaufman, 69, surrendered herself on Friday to authorities in San Francisco . She was found guilty by a jury for her role in an action last year that saw members of the group StopAI chain and lock the front doors of OpenAI's headquarters in protest against the pursuit of artificial superintelligence. The retired teacher from Berkeley, California, refused to move from a sit-in protest in February 2025 and pleaded not guilty to multiple misdemeanor charges. She was convicted in June of interfering with a business, trespassing with intent to interfere with a business, unlawful assembly and refusal to disburse a riot.

Apple trains its own AI model for China market with Alibaba's support, sources say People walk past a booth showcasing Apple's suppliers during the China International Supply Chain Expo in Beijing on June 22. Apple has trained a large language model specifically for the China market, three people familiar with the matter said, a departure from the iPhone maker's strategy of relying on third-party models to power AI features in the country. The AI model was developed in partnership with Alibaba Group and trained with the Chinese tech giant's support, the people said, declining to be named as the information is sensitive and not public. Apple's training of its own China-specific model has not been reported before. The company had previously leaned toward domestic partners' models to bring generative AI to iPhones and other devices sold in China, where U.S. models such as Anthropic's Claude and OpenAI's ChatGPT, which Apple pairs with its own technology in its home market, are not available.