Senate Commerce Chair Ted Cruz said in an interview Thursday he is looking to take up legislation to address the “catastrophic risk” posed by artificial intelligence — increasingly top of mind after a major OpenAI security incident this week.
It’s a first glimpse at the types of bills that could be on the agenda for a big markup of AI and kids’ online safety measures the Texas Republican plans to hold the first week of August, and about which he has been largely secretive.
“Look, it’s deeply concerning, and there are undoubted risks and catastrophic risks that that AI potentially brings, and how the government deals with those risks and minimizes the chance of harm is a difficult policy question that Congress is wrestling with,” Cruz said.
OpenAI earlier this week disclosed that its most powerful model available to the public, and another that has yet to be released, were both able to break free of a controlled laboratory test and hack into another AI developer platform — the first documented caseof a fully autonomous AI cyberattack.
Cruz said Thursday that legislation confronting this scenario is “among the topics we are looking at for the markup,” but declined to say whether he was writing a bill to address it — or if he was working across the aisle on a bipartisan proposal.
“More to come,” he said. “We are negotiating with the Democrats and deciding — trying to find the bills that have the best maximum positive impact and also best chance of bipartisan support and passing law.”
Cruz’s aides, granted anonymity to speak freely, have previously said that the Commerce Committee chair believes in very targeted federal action to address issues like catastrophic risk or deepfakes — rather than a massive regulatory framework for AI.
The Senate Commerce markup on AI legislation was supposed to happen the last week of July but was pushed back amid continued jockeying over which bills would be included on the docket.
There has, in particular, been confusion over the fate of a forthcoming package from Sen. Marsha Blackburn (R-Tenn.), who has been working with the White House to pair a kids’ online safety measure with provisions to preempt state AI laws.
That effort suffered a blow this week when President Donald Trump declined to back the effort. Cruz, on his podcast Thursday, took credit for that decision, saying he announced during a White House meeting with Trump and Blackburn that Blackburn’s emerging framework would not have the votes to pass the Senate.
Trump was ready to hit “send” on a Truth Social post endorsing the Blackburn proposal before Cruz interjected, the Texas senator said.
Blackburn’s office did not return a request for comment about Cruz’s version of events, and Cruz said Thursday in an interview he has not spoken to Blackburn since their visit to the Oval Office.
Cruz could also include in his markup a new bill he introduced Thursday with Sen. Raphael Warnock (D-Ga.) that would create a commission to study AI developments — focusing on bioethics — and report the findings to Congress.
Read the full article here