The Republican cosponsor of bipartisan artificial intelligence safety legislation said he sat down with OpenAI’s top lobbyist a day before the company backed a key provision in the bill that’s struggled to gain traction on Capitol Hill.
“We talked about the urgent need for action,” Rep. Jay Obernolte of California told reporters of his Monday meeting with OpenAI chief global affairs officer Chris Lehane. “He was actually asking how he could be helpful. So, we’re trying to craft a path forward.”
Obernolte continued, “I told him that we all needed to get on the same page and push in the same direction, all pushing together. That’s the only way we’re going to be able to get something done this year.”
Obernolte has worked for months with Rep. Lori Trahan (D-Mass.) on a comprehensive package that would, among other things, enable third-party evaluators to examine AI models to ensure they comply with safety standards. Lehane told reporters Tuesday that OpenAI would support that specific provision.
It comes at a pivotal moment in the debate over what role Congress should play in regulating the AI industry after the CEOs of leading AI labs, including Sam Altman of OpenAI, issued warnings that without new guardrails the rapidly evolving technology could do irreparable harm to humanity in the very near future.
Obernolte and Trahan’s work on an AI safety bill long preceded the current conversations happening in Washington, but both lawmakers are now making the case that their measure should be the starting point for legislative action.
Trahan told reporters Tuesday it was “good to hear” that a component of their bill now had OpenAI’s support. She said it was a sign the legislation was building the necessary coalitions of support after she and Obernolte consulted with a broad cross-section of industry players to draft their package.
She also underscored the need for Congress to set parameters for AI use and development even as some Republicans, and President Donald Trump, insist the federal government should not interfere with private-sector innovation.
While some AI companies, including Anthropic, have supported letting evaluators into their company to scrutinize their models, Trahan said such agreements should be mandatory — not voluntary.
“It’s not just up to industry to decide, despite what the president might say,” she said from the sidelines of the Pro-Human Assembly, a gathering of high-profile speakers expounding on both the risks and opportunities of AI. “There are a whole lot of stakeholders who deserve a voice at the table.”
“Congress needs to take the next step,” Trahan added.
Read the full article here