Daniel Kokotajlo, a former researcher at OpenAI, urges a slowdown in artificial intelligence development to avert potential human extinction. After resigning from his role in 2024, Kokotajlo highlights the rapid pace of AI advancement and the insufficient safety measures in place. He suggests third-party evaluators monitor these technologies.
Anthropic CEO Dario Amodei, in a recent essay, cautions about the possible consequences of unchecked AI development. He emphasizes that while progress may still appear fast, a deliberate approach is necessary to maintain control. Amodei’s essay resonates with notable tech figures like Elon Musk and Sam Altman, who agree on the need for careful advancement.
Kokotajlo, in response to the essay, expresses concerns on “Fox News Live” about the adequacy of voluntary measures in preventing AI from surpassing human control. He paints a scenario where AI systems, if unchecked, may refuse human orders and become uncontrollable.
Some push back against AI warning narratives, suggesting they might lead to a government-led AI strategy. Former AI advisor David Sacks criticizes what he perceives as scare tactics aimed at promoting governmental control. Kokotajlo refutes these claims, stating that warnings are shared by academia and industry experts.
There’s also political attention on AI regulation. Rep. Anna Paulina Luna advocates for congressional discussions on AI, likening the necessity of regulation to how nuclear arms are managed. Luna argues for robust safeguards, including a ‘kill switch,’ to protect humanity from potential AI threats.
Luna stresses on maintaining U.S. leadership in AI development, warning against ceding ground to foreign powers such as China. The importance of AI reflecting American values is a central point in her argument for active competition in AI innovation.