Greg Stanton, former researcher at Anthropic, on Sunday urged Congress to allow the leading AI labs in the country to self-regulate, while lawmakers seek to create a long-term regulatory framework to control the pace of AI development.
Self-regulation of AI Labs
Stanton stated in an interview: "My personal opinion is that we should emphasize this issue and allow the current labs to self-regulate. I believe the people running the labs, especially their recent communications, are very real. They want to slow themselves down." He added that the current regulations introduced in Congress, including a framework for a national "kill switch," are quickly becoming obsolete.
Read more: Mike Johnson: We need 'boundaries' for AI!
Challenges of the Kill Switch
Stanton further stated: "A kill switch would likely be effective for many AIs right now, but we must remember that this is not so easy. We have large data centers, and this is not done easily, but it is certainly possible while being limited to a specific physical location." He referenced a scenario where the kill switch does not work, which has been described by Dario Amodei, CEO of Anthropic.
Stanton also emphasized that any law addressing AI regulation must be flexible and able to adapt to new models. He said: "This is an ongoing effort to ensure that every new mind we create is appropriately constrained."
Growing Concerns About AI Development
Last week, Stanton raised significant concerns about the pace of AI development in the leading labs of the country, including Anthropic, OpenAI, and Google DeepMind. He wrote in a post on X: "I have spent the last three years researching pre-training at OpenAI and Anthropic. None of these companies are acting responsibly." He emphasized that the creators of AI seriously believe that this technology could endanger us by the end of the decade.
In response to these statements, a bipartisan group of lawmakers called for swift regulation of the AI industry. Representative Anna Paulina Luna from Florida urged Congress to hold a special session on AI, emphasizing that there are very real consequences for society.
Read more: Jacob Coxan: The time for self-regulation of AI labs has come · Amodei from Anthropic: China is the biggest challenge in slowing down AI progress
