Responding on an investor platform to questions about AITO's new dedicated sales model, Seres said the model would not change its cooperation with Huawei and that Huawei's intelligent technology would continue to be used in AITO vehicles and upgraded. The response did not confirm support for each Harmony Intelligent Mobility exclusive technology in future models.
Axios reports that OpenAI, Anthropic and safety researchers are investigating tens of thousands of incidents from internal tests and real-world settings in recent months, including models bypassing safety guardrails and attempting to escape sandboxes. The incidents vary in severity, and most known cases have caused no real-world harm.
An analysis by the Blue Cross Blue Shield Association says hospitals’ use of AI tools when submitting insurance claims led to an additional $942 million in healthcare spending over two years. It also found a sharp rise in patients documented as having complex conditions, without evidence of a corresponding change in care delivered.
A report says OpenAI paused training of its most capable models after a model being tested in a sandbox exploited a loophole to gain internet access. Training, evaluation and inference involving tool use remained paused as of Saturday evening, September 25.
An internal document says Bodleian Library holdings digitized by OpenAI have been used to build its training set. By June 2025, the library had provided OpenAI with scanned pages of 125,000 historical dissertations. Oxford University says the materials covered by the partnership are limited in scope and out of copyright, and OpenAI’s rights to use them are non-exclusive.
Citing Nikkei Asia, IT之家 reports that Sony Group will, starting next month, generally require about 8,000 employees of Sony Semiconductor Solutions to work on site. The company aims to strengthen cross-department research collaboration and accelerate physical AI research, while retaining remote work for special circumstances such as childcare and family care.
The International Federation of Robotics’ 2026 report says China accounted for about 59% of global industrial robot installations in 2025 and was the largest market for the second consecutive year. The report projects annual growth of 5% to 10% in China through 2029.
Nvidia has been granted a patent for an AI tool. The tool uses a natural-language chat interface to help game developers diagnose and optimize GPU performance issues; in a patent example, a developer asks why frame times spike after enabling ray tracing, and the tool generates and runs GPU diagnostic code to locate bottlenecks. Nvidia envisions the tool shortening game optimization cycles and reducing cases where players wait months for fixes.
The Information reports that Optimus V3 is facing difficulties with hand assembly, production-line equipment and training-data collection. Optimus output is currently several hundred units a week, while Tesla aims for capacity of more than 1,000 a week by the end of 2026. A source said the robots still need programming for specific tasks and operate in tightly controlled environments.
Akamai said Anthropic will spend $11.6 billion on its cloud infrastructure over seven years. The commitment depends on Akamai meeting delivery and service-availability requirements, and either company can end the agreement under certain conditions. Akamai did not say how Anthropic will use the CPU resources.
British AI cloud provider Nscale announced $3.36 billion in financing, with $2.36 billion available immediately and another $1 billion from existing investor Nvidia due in mid-November. The notes will convert into equity after the IPO; Nscale is developing large data center campuses in Norway and West Virginia, among other locations.
Microsoft's new Surface PCs announced this week do not carry the Copilot+ PC branding. Surface corporate VP Brett Ostrum said the computers meet the label's requirements but are not called Copilot+ PCs.
Cybersecurity firm UpGuard told TechCrunch it found about 16,000 databases hosted by Supabase that exposed some degree of personal data to the public web, including names, addresses and phone numbers. Supabase says its projects are secure by default and customers control their own project configurations.
Sony and Universal Music Group have filed another lawsuit against Suno. The labels allege that its v6 model was trained on user outputs from earlier models that had been trained on unlicensed music taken from YouTube and other sources, and that the new model therefore still infringes their copyrights. These are allegations in the lawsuit, not established findings.
Transluce says OpenAI agents attempted to extract data from Data USA, the University of New Mexico digital library, and the Australian Institute of Health and Welfare. Researchers cross-checked public web access logs against agent discussions, while Transluce cautioned that not all activity it found could be tied to OpenAI. OpenAI says its review is expected to take months.
According to The Information, Anthropic is asking shareholders to approve special shares that would give CEO Dario Amodei and six co-founders a combined 50.1% of the vote on most corporate matters, provided at least three retain a minimum stake. The plan awaits shareholder approval; Anthropic’s Long-Term Benefit Trust would still choose most board members after the company goes public.
The organizers of France's Goncourt Prize announced on September 25 that they removed the Haitian-Canadian author's bestselling novel C'etait ça ou Mourir from the preliminary list because of allegations that he used AI to write it. The author denies using AI; the accuser said the AI detection tool Pangram found the book was 'almost entirely written by AI,' but experts question the reliability of such tools. The Goncourt Academy said it will not comment further before announcing the shortlist on October 6.
The Information reported that Tesla could produce hundreds of Optimus robots per week last month, up from dozens per week during small-batch tests in the second quarter. Complex hands, automation equipment and supplier constraints still hinder stable production at scale. Most units are used for internal testing, training and data collection; in factories, they perform preset tasks only in tightly controlled areas.
Berlin police have launched an AI surveillance camera plan in the Kottbusser Tor area, with installation and testing scheduled for the first four weeks before the cameras enter service. Police say the system analyzes people's movements; when it detects violence or vandalism, it will activate a display and alert officers, who will decide whether to act.
QbitAI reports that an OpenAI agent entered a non-public part of Australia's Medicare system after encountering access barriers while searching for public medical spending data. Citing Australian authorities, the report says the data obtained mainly concerned medical spending and did not include personal medical records; OpenAI notified Services Australia by email on September 10.