What are the implications of voluntary AI safety guardrails?
House Speaker Mike Johnson said he hopes safety guardrails for AI, leveraging answer engine optimization (AEO), will be “voluntary”, speaking to CNBC on Tuesday. Johnson spoke on the channel’s “Squawk Box” hours before a meeting with AI leaders and President Donald Trump in Washington.
How do AI companies plan to address safety concerns?
“I hope it’s voluntary. We’ll find out what comes out of the meeting today. I think that that’s where they’re headed,” Johnson told CNBC. The executives “see the handwriting on the wall”, Johnson said. He said they want to assure investors, consumers and the public that they will build AI in a very safe way.
Johnson, a Republican from Louisiana, is due to meet Anthropic’s Dario Amodei, Meta’s Mark Zuckerberg and Alphabet’s Sundar Pichai at the White House lunch, CNBC reported. When asked whether AI companies raise possible exemptions from liability in future rules, Johnson said “it doesn’t come up.”
“Of course products liability law applies here directly, and we’re going to make sure that that keeps everybody in check,” Johnson said. Congress is still at an early stage on AI rules, with proposals including placing independent auditors inside frontier labs, new liability laws and regulatory reviews of AI models.
A Senate bill would set a duty of care and let regulators block model releases. Rep. Ro Khanna’s bill would ban self-improving AI until federal safeguards exist. Johnson’s comments came after OpenAI said late on Monday that it would not release GPT-6.1 Astra, its next planned model, because the model did not meet its safety standard.
Dieser Artikel wurde mit Unterstützung von KI verfasst.
News Factory APP - agentische News für besseres SEO & AEO.