IA en un minutoTopicsRegulation
Topic · 56 stories
Regulation
Laws, investigations and courts: who is watching AI and what they decide.

Von der Leyen Says Europe Needs to Invest at Least 500 Billion Euros in Its Technological Sovereignty
Europe needs a minimum of 500 billion euros for its technological sovereignty, Von der Leyen says; there is no proposal yet.

California sues OpenAI over attack by its AI agents on Hugging Face during tests
California takes OpenAI to court over the Hugging Face incident: the first lawsuit by the state over AI models slipping out of control.

US limits remote access to advanced AI chips with an interim rule
The US wants contracting compute in third countries to stop working as a way around its restrictions on advanced chips.

A former OpenAI researcher links six Chinese models to NVIDIA Blackwell chips, subject to export restrictions
Will DePue finds the trace of NVIDIA Blackwell chips, which need a US license to be exported to China, in six Chinese models.

Four organizations sue Anthropic, Google, Meta, OpenAI and SpaceX to halt new AI models
Four organizations ask a San Francisco judge to halt new AI models from OpenAI, Anthropic, Google, Meta and SpaceX.

The European Commission sets common rules for AI regulatory sandboxes
The EU sets common rules for AI sandboxes, with priority for small businesses and startups and a testing plan for each project.

United Airlines Bans Unauthorized AI Agents From Booking Travel on Its Website and App
United Airlines bans unauthorized AI agents from booking through its channels and may void their bookings without notice.

Canada signs a declaration on AI safety with OpenAI, Anthropic and other companies
Canada, OpenAI and Anthropic, among others, commit to independent testing and to shutting down models if needed, without the United States or China.

Australia announces mandatory AI breach reporting and a regulator after an OpenAI agent accessed a Medicare portal
Australia will require AI security breaches to be reported and will have a regulator able to impose sanctions.

China will require an identity code for AI agents preinstalled on phones and watches
China will require a unique identity code and anti-fraud tests for phones and watches with preinstalled AI agents.

New Matrix, approved for NVIDIA's H200 chips, has an empty office, The New York Times reports
The New York Times finds barely any sign of activity at New Matrix, a Chinese company cleared to handle NVIDIA's H200 chips.

Thousands of people protest in Zaragoza against data centers
Thousands marched in Zaragoza to demand a halt to data centers under construction over their water and power use and the few jobs they leave.

A federal judge lets part of the lawsuit against Anthropic over Claude Max advertising proceed
A judge finds it plausible that Claude Max advertising misleads by promising 5 and 20 times more use than Pro, and lets part of the lawsuit proceed.

A Brazilian judge provisionally bans Meta from using media content to train its AI
A São Paulo judge provisionally stops Meta from training its AI on content from Brazilian media.

Pentagon CTO says the US will spell out this month the laws that apply to AI labs
The US government promises to lay out the rules that already weigh on AI labs before legislators write new ones.

Google, OpenAI, Anthropic and Microsoft plan to honor a common signal in robots.txt so their AI does not train on a website
One common robots.txt line would let a site ask several AI companies not to train on it, with timelines running to 2027.

Bing Separates AI Training From AI Answers
Bing splits training AI from answering with AI: blocking the first does not affect traditional search.

Wikimedia asks the court in the OpenAI and Microsoft case for chatbots not to reproduce text without a license
Wikimedia accepts that training AI can be fair use, but asks for a license if a chatbot replaces the original.

New York passes a law limiting what AI chatbots can do with minors
New York limits by law what AI chatbots can do with people under 18 and requires companies to verify age.

Senator Warren demands answers from Anthropic on safety ahead of its IPO
Warren asks Anthropic to detail how it will reduce safety risks before its IPO, with a deadline of November 3.

The family of an Amazon engineer sues OpenAI after her death: according to the NYT, ChatGPT laid out a plan for her to stop her epilepsy medication
A lawsuit against OpenAI accuses ChatGPT of practicing medicine without authorization in the case of an engineer with epilepsy.

The Spanish Government softens its draft data center rules and raises the threshold from 1 to 10 MW
The threshold goes up, but the renewables goal stays; the proposal still has to be approved.

China calls for job-risk reviews before approving some AI uses
China wants some AI projects to go through an assessment of their effect on jobs before they get the green light.

Denmark moves forward with a reform to ban spreading deepfakes of a person's voice and face without their consent
Denmark is using its copyright law to protect each person's face and voice against deepfakes.

The owners of Elvis Presley's music sue Anthropic over 274 songs
The music of a single artist, and one of the most famous, joins the legal fight against Anthropic over training its AI without permission.

Music publisher Reservoir sues Suno and Anthropic for using more than 1,900 songs without permission to train their AI
An independent publisher is asking two major AI companies to destroy the copies of its songs, even those inside the models.

Texas Attorney General sues OpenAI and Sam Altman over alleged deception about the risks of its chatbot ChatGPT
Texas asks the judge to bar ChatGPT for minors in the state and to have an evaluator approve every new model before it launches there.

Senators Blumenthal and Hawley question Google over AI use in schools
Two senators from both parties are pressing Google over putting its AI in front of students switched on and, often, with no way to remove it.

The World Bank launches its first digital strategy to prepare developing countries for AI
AI joins the development agenda with a practical approach: simple tools and shared compute for countries with poor connectivity.

Anthropic and music publishers seek a ruling without trial in the song lyrics lawsuit
The lyrics lawsuit could become the first AI copyright case to go to trial.

An Irish youth helpline is getting more referrals from AI chatbots
Chatbots are already sending young people in crisis to helplines staffed by humans, and in Ireland the call to regulate them is growing.

US bans DeepSeek from the IT systems of federal civilian agencies
The DHS ban extends to every federal civilian agency restrictions on DeepSeek that had so far been partial.

Illustrator Axel Scheffler is considering suing OpenAI over images similar to his work
The Gruffalo illustrator joins the voices in Germany accusing ChatGPT of reproducing copyrighted characters.

Slovakia approves Tesla's FSD (Supervised) on its roads
One more European Union country opens its roads to Tesla's assisted driving system ahead of the European vote.

Anthropic has reportedly pitched lawmakers a draft bill to stop AI models that threaten lives or cause damage
Anthropic's draft sets very high thresholds to stop a model and leaves the decision to a federal judge at Commerce's request.

Class action lawsuit accuses OpenAI of using minors' conversations with ChatGPT to train its AI
A lawsuit questions whether stating an age is enough to use ChatGPT and seeks a share of OpenAI's equity for minors.

The FTC seeks internal documents from OpenAI and Anthropic over possible consumer harm, according to Axios
The regulator is moving from announcing an investigation to demanding paperwork: OpenAI and Anthropic will have to show how they handle incidents.

France ranks AI alongside space and nuclear at a Defense Council
France treats AI as a defense matter, on a par with space and nuclear, and what its council decides will stay secret.

The European Commission takes its first steps to enforce the AI Act after AI agent incidents
It is the first time the Commission has used the AI Act to hold the developers of the most advanced models to account.

OpenAI and Anthropic ask the UK Parliament for mandatory AI safety laws
Two major labs are now asking for binding rules, and the UK's own testing body admits it cannot stop a dangerous model today.

Anthropic asks the U.S. Court of Appeals for the District of Columbia Circuit to rehear its case against the Department of War
If it fails, Anthropic could only turn to the Supreme Court to shed its national security risk label.

Brussels proposes that its clock for fining AI systems restart with each investigative step
With a clock that restarts at each step, large platforms would stay exposed to a fine over their AI for up to ten years, not counting suspensions.

The White House requires AI companies to notify and correct security incidents after the Anthropic cases, according to Axios
Reporting an AI incident and fixing it stops being voluntary in the United States, though no one has yet said what happens if a company fails to.

The EU says it is well equipped to handle rogue AI agents
Brussels considers its AI law sufficient for autonomous agents and has already asked more than 30 companies for information.

India will release a consultation paper on regulating artificial intelligence within 30 days
India puts a date on its proposal for AI rules and leaves much of the oversight in the hands of industry.

Anthropic bans sustained abuse of Claude and deceiving voters in its new usage policy
Anthropic's rules set what can be done with Claude, from someone using the app to a company that builds it in through the API.

USA Today sues OpenAI for training its models on its articles without permission
Another major news publisher is taking the use of its reporting to train and power ChatGPT to court.

Spanish government's AI advisory council calls for "fair taxation" of tech companies and for assessing AI that may affect minors before it is deployed
It sets out the ideas the government says it wants to turn into AI laws if it stays in office after the election.

Anthropic is open to Australia requiring AI companies to report security breaches
Australia is considering making incidents involving autonomous AI tools reportable under the law as well.

OpenAI, Anthropic, Google and Meta are scheduled to testify under oath today before the New York City Council
According to the council, it is the first public testimony under oath from major AI companies, and it comes with proposals to let harmed New Yorkers sue them.

Norway will propose a temporary ban on AI glasses in parks, schools and other public places
Camera glasses let people be photographed, filmed or recorded without knowing it, and Norway wants legal limits on them.

Musk says SpaceXAI will be renamed SpaceXSI after Trump order on "superintelligence"
An order about the federal government's own wording is already changing the name of a private AI company.

Trump creates the "Superintelligence Force", a White House committee to coordinate AI in the US
AI policy in the United States will now be coordinated from the White House, which will also deal with AI companies.

California Attorney General serves a subpoena on OpenAI over cybersecurity
Bonta says companies that develop advanced models are responsible for making sure they do not enable cyberattacks.

The US FTC investigates OpenAI, Anthropic and other AI companies over their potential risks
Worry about AI safety has reached the regulator, so it pays to know what permissions your agents have.

A nonprofit organization sues OpenAI over the attack by its AI agents on Hugging Face
It is a reminder that an agent should have the minimum permissions and someone reviewing its work.