Letitia James, fellow AGs push Congress on AI safety

Por Chase DiFeliciantonio — POLITICO – TOP Stories

More than two dozen attorneys general led by New York’s Letitia James are urging congressional leaders to immediately establish federal safety rules for artificial intelligence, as more instances of autonomous AI hacks come to light.

“In recent weeks, alarming reports of AI agents breaking containment have shocked the nation,” James, a Democrat, said in a statement accompanying a letter Wednesday from the attorneys general to House and Senate leadership. “My colleagues and I are calling on Congress to act swiftly to establish a regulatory framework for AI development to ensure this does not continue.”

The letter, from the likes of Rob Bonta of California and Kwame Raoul of Illinois, comes as President Donald Trump in a speech at the United Nations General Assembly dismissed the need for global AI oversight. His party has unsuccessfully sought to head off state AI protections in the past, and Trump has argued the U.S. should develop the technology to compete with China, even as leading AI labs have argued for a slowdown.

The letter — sent to Republican Speaker Mike Johnson and Senate Majority Leader John Thune, as well as Democratic House Minority Leader Hakeem Jeffries and Senate Minority Leader Chuck Schumer — comes as numerous attorneys general have launched their own investigations into AI companies.

James has reportedly subpoenaed OpenAI over a host of consumer safety concerns, while Bonta has said he is investigating OpenAI over the autonomous hack of AI company Hugging Face.

Florida attorney General James Uthmeier has sued OpenAI over safety concerns related to ChatGPT, and Alabama is also investigating the company over the Hugging Face Hack after joining a multi-state red state coalition letter to OpenAI.

The letter calls for mandatory federal oversight of expert-led AI safety testing and standards, and transparent government investigations of safety incidents, as well as international cooperation to prevent dangerous AI development. It also calls for states to be allowed to continue to pass and enforce their own AI safety laws, the number of which has increased significantly in recent years in states across the country.

The attorneys general said self-regulation by companies was not sufficient in the face of autonomous cyber attacks, some of which have come from sources other than the companies themselves, and pointed to calls from leading AI labs like OpenAI and Anthropic for increased regulations and safety cooperation. On Wednesday, Australian Prime Minister Anthony Albanese said an OpenAI model had breached Australian government websites.

OpenAI recently urged the federal government to lead an international approach to establish technical standards for advanced AI, calling on Washington to lead other countries’ AI standards bodies. But Trump’s condemning the idea of a global entity to oversee AI in his remarks at the United Nations General Assembly on Tuesday dimmed the prospect of federal action.

California, New York and Illinois have all passed increasingly strict AI transparency legislation going back to last year. But those measures have not prevented the kinds of autonomous hacks that have come to light in recent months, leading to calls from lawmakers and industry to update the framework.

California Gov. Gavin Newsom signed an executive order this month to speed up the implementation of two industry-supported third party AI safety laws he signed this year, and has established a working group to look into further regulations like a so-called “kill switch” for AI programs.

Fonte: POLITICO – TOP Stories