Yann LeCun Urges Signing Letter to Block Regulation for AI Research

Computer scientist Yann LeCun has opened up on the proposed 'AI Safety Bill' under consideration in the California Legislature that could force some of the biggest AI companies to change the way they do business.

“Regulating research and development would have apocalyptic consequences on the AI ecosystem,” the Meta scientist said in a post on LinkedIn. LeCun has signed a letter calling for stopping this bill and has asked others to do the same.

What Is The Bill About

Last month, the California State Senate passed legislation to control the development and training of complex, cutting-edge AI models, with the goal of preventing bad actors from using them for evil ends.

The AI Safety Bill mandates AI developers to 'include a kill switch that can shut down AI systems to prevent AI models from inflicting serious harm.'

It would also compel AI developers to report their compliance efforts to the 'Frontier Model Division,' which will be established within the California Department of Technology. Companies that fail to comply with these rules may face lawsuits and hefty penalties.

Supporters of the bill say that quickly expanding technology necessitates safeguards, and the law is a reasonable—if restricted and imperfect—first step towards legally enshrining safety best practices.

Silicon Valley Fumes

The passing of the bill has sparked outrage among developers, many of whom operate outside of the state, claiming that it will severely impede innovation.

Among the worst criticisms is that the bill will affect open-source AI models, in which creators make their source code freely available to the public, letting developers build on top of it, such as Meta's flagship LLM, Llama. The law would hold open model makers potentially accountable for bad actors who use their models to hurt others.

Clement Delangue, CEO of HuggingFace, described the law as a "huge blow to both California and US innovation."

"If someone wanted to come up with regulations to stifle innovation, one could hardly do better," said Andrew Ng, a renowned computer scientist who pioneered AI programmes at Alphabet's Google and China's Baidu and currently serves on Amazon's board.

Arun Rao, Meta's lead product manager for generative AI, wrote on X last week that the law was "unworkable" and would "end open source in [California]".

Several major IT trade associations have also spoken out against the bill.

In Support

Geoffrey Hinton and Yoshua Bengio, who have spoken out about the potential existential concerns posed by AI, support the AI Safety Bill.

According to Hinton, the proposal "takes a very sensible approach to balance those concerns."

Way Forward

The California AI Safety Bill discussion underscores the complexities and divisions that come with regulating developing technology such as AI. As the sector evolves, policymakers must strike a difficult balance between encouraging innovation and guaranteeing the appropriate development of AI systems.