EU Adds ChatGPT to Oversight List as US, China Chart AI Paths

The US, China and the EU are each charting their own course for regulating artificial intelligence. The US under President Donald Trump has scrapped earlier safety rules and leans on voluntary industry pledges. China already has evolving rules and stresses keeping AI under human control. The EU adopted the AI Act in 2024 and added ChatGPT to its digital services oversight list, aiming to shape global standards. All three regions face pressure to keep AI safe as the technology advances quickly.

Rep. Lori Trahan, a Democrat from Massachusetts, released a discussion draft of legislation focused on AI liability. The draft proposes holding companies accountable for the actions of their AI systems and would create a new federal agency to oversee AI development. Some in the tech industry worry the rules could slow innovation.

The White House is replacing the term AI with Super Intelligence, or SI. SI is currently defined the same way as AI, but officials have 60 days to create a formal federal definition. A recent White House summit brought together chipmakers, data centers, cloud platforms and frontier model companies. Business leaders are urged to treat SI as a strategic term for planning and innovation.

SpaceX unveiled a plan to put a million satellites in orbit to power AI data centres. The visualisation was shown at the 2026 International Astronautical Congress in Turkiye. The project aims to bring AI computing capabilities to space, marking a bold step in how AI infrastructure is built.

Stanford neuroscientist Andrew Huberman says Meta, OpenAI and Anthropic are really chasing biotech goals like non-invasive brain read/write. He argues large language models are just a waypoint, not the final destination. He expects non-invasive stimulation hardware to move from sleep into waking focus tools within 7 to 12 months.

Health Secretary Robert F. Kennedy Jr. backed AI at a MAHA summit, saying it will let people fact check their own medical advice. OpenAI and Anthropic sponsored the summit, along with many other companies with business before the government. Some MAHA allies criticized the move, saying the movement is getting too close to corporate interests. Medical experts like Dr. Zeke Emanuel say Kennedy misunderstands how AI works.

AI is changing how people search for apartments through chatbots, AI assistants and rental platform tools. Companies like Wilbur Properties, RentCafe and Zumper are using or testing AI systems, including Anthropic's technology and assistants named Ren and Zoe. California's Department of Real Estate warns that real estate professionals must still review AI-generated information.

South Korea is facing a new wave of cyberattacks that use artificial intelligence. Multiple financial institutions have reported service disruptions and suspected data breaches. In one incident an AI-assisted system reportedly found vulnerabilities and compromised data for 25,000 customers. Banks and regulators now need stronger defenses to detect and respond to AI-assisted threats.

Buyers of AI-enabled companies may inherit liabilities they think are covered by insurance. Standard insurance forms from the Insurance Services Office now exclude many generative AI losses, and W.R. Berkley added an absolute AI exclusion across several policy types. Routine acquisition checks often miss these changes because they focus on whether policies exist rather than what they cover.

Mumsnet denied using AI to write content after a detailed prompt appeared on its forum. The prompt, titled AIBU Response Schema v.1.0, gave instructions for writing a response post. Founder Justine Roberts said the site uses AI for coding and moderation but not for writing posts.

Key Takeaways

The US, China and EU each take different paths to regulate AI, with the US relying on voluntary pledges, China emphasizing human control, and the EU adopting the AI Act in 2024.
Rep. Lori Trahan released a discussion draft proposing AI liability rules, a new federal agency and strict accountability for companies using AI systems.
The White House is replacing the term AI with Super Intelligence (SI) and has 60 days to create a formal federal definition.
SpaceX revealed a plan to deploy a million satellites in orbit to power AI data centres, unveiled at the 2026 International Astronautical Congress in Turkiye.
Andrew Huberman says Meta, OpenAI and Anthropic are pursuing biotech goals like non-invasive brain read/write, with hardware expected in waking focus tools within 7 to 12 months.
RFK Jr. backed AI at a MAHA summit, while OpenAI and Anthropic sponsored the event, drawing criticism from allies and medical experts.
Rental platforms like Wilbur Properties, RentCafe and Zumper are using or testing AI tools including Anthropic's technology and assistants named Ren and Zoe.
South Korea reported AI-assisted cyberattacks on financial institutions, with one breach compromising data for 25,000 customers.
Insurance forms now exclude many generative AI losses, and W.R. Berkley added an absolute AI exclusion, leaving buyers of AI-enabled companies exposed.
Mumsnet denied using AI for writing posts after a prompt appeared on its forum, saying AI is used only for coding and moderation.

US China EU take different paths to regulate AI

The United States China and the European Union are each taking their own approach to governing artificial intelligence. The US relies on voluntary pledges and executive orders under President Donald Trump. China already has evolving rules and emphasizes keeping AI under human control. The EU adopted the AI Act in 2024 and aims to shape global standards. All three face pressure to keep AI safe as the technology advances rapidly.

US China EU take different paths to regulate AI

The US China and the EU are following separate paths to control artificial intelligence. In the US President Donald Trump has scrapped earlier safety rules and relies on voluntary industry pledges. China has complex rules and warns that hostile forces use AI to spread harmful information. The EU adopted the AI Act in 2024 and added ChatGPT to its digital services oversight list. All three regions are working to balance innovation with safety.

Mumsnet denies using AI for content after prompt appears

Mumsnet denied using AI to write content after a detailed prompt appeared on its forum. The prompt titled AIBU Response Schema v.1.0 gave instructions for writing a response post. Users called the appearance really creepy and demanded an explanation. Mumsnet first said the prompt came from a title suggestion tool then said it came from a user's computer. Founder Justine Roberts said the site uses AI for coding and moderation but not for writing posts.

Insurance Gaps Mean Tech Acquisitions Can Inherit AI Liabilities

Buyers of AI-enabled companies may inherit liabilities that they believe are covered by insurance. Standard insurance forms from the Insurance Services Office now exclude many generative AI losses. These exclusions appeared in forms dated January 2026 including CG 40 47 CG 40 48 and CG 35 08. W.R. Berkley also added an absolute AI exclusion across several policy types. The problem is that routine acquisition checks often miss these changes because they focus on whether policies exist rather than what they cover.

AI Agents pose growing cyber threat to South Korea banks

South Korea is facing a new wave of cyberattacks that use artificial intelligence. Multiple financial institutions have reported service disruptions and suspected data breaches. In one incident an AI-assisted system reportedly found vulnerabilities and compromised data for 25000 customers. AI can automate tasks like scanning websites and generating malicious code making attacks faster and more sophisticated. Banks and regulators now need stronger defenses to detect and respond to AI-assisted threats.

Huberman says AI labs are really biotech

Stanford neuroscientist Andrew Huberman says Meta, OpenAI and Anthropic are really chasing biotech goals like non-invasive brain read/write. He argues large language models are just a waypoint, not the final destination. He also discussed sleep protocols and a new MIT-linked eye mask that measures REM and uses gentle ear stimulation. He expects non-invasive stimulation hardware to move from sleep into waking focus tools within 7 to 12 months.

RFK Jr.'s embrace of AI confounds some MAHA allies and medical experts

Health Secretary Robert F. Kennedy Jr. backed AI at a MAHA summit, saying it will let people fact check their own medical advice. Some MAHA allies criticized the move, saying the movement is getting too close to corporate interests. Medical experts like Dr. Zeke Emanuel say Kennedy misunderstands how AI works. OpenAI and Anthropic sponsored the summit, along with many other companies with business before the government.

Trahan unveils AI liability discussion draft

Rep. Lori Trahan, a Democrat from Massachusetts, released a discussion draft of legislation about AI liability. The draft proposes a new framework that would hold companies accountable for the actions of their AI systems. It would create a new federal agency to oversee AI development and impose strict rules on companies using AI. Some in the tech industry are skeptical and say it could slow innovation.

How AI is changing the apartment search

AI is changing how people search for apartments through chatbots, AI assistants and rental platform tools. Companies like Wilbur Properties, RentCafe and Zumper are using or testing AI systems such as Anthropic's technology and assistants named Ren and Zoe. California's Department of Real Estate warns that real estate professionals must still review AI-generated information. A renter in Palo Alto used Claude and ChatGPT to search but ultimately found a place through word of mouth.

AI Has a New Name: Super Intelligence

The White House is replacing the term AI with Super Intelligence, or SI. SI is currently defined the same way as AI, with 60 days to create a formal federal definition. A White House summit brought together chipmakers, data centers, cloud platforms and frontier model companies. Business leaders are urged to treat SI as a strategic term for planning and innovation.

SpaceX reveals plan for million-satellite AI data centres in orbit

SpaceX showed a new visualisation of its plan to put a million satellites in space. The satellites would power AI data centres in orbit. The plan was unveiled at the 2026 International Astronautical Congress in Turkiye. The project aims to bring AI computing capabilities to space.