El-Sayed backs AI kill switch, safety agency

A proposal aimed at bringing artificial intelligence development under democratic control has been put forward, advocating for significant changes in how powerful AI technologies are governed. The plan, outlined by a US Senate candidate, emphasizes public interest and control over profit-driven innovation.

The proposal includes several key pillars. Firstly, it calls for all "frontier AI labs" to be chartered as public benefit corporations, legally obligating them to balance public interest with profit motives. This charter would embed a "First, Do No Harm" principle, ensuring that public safety, civil rights, and democratic integrity are not compromised for financial gain.

Furthermore, the plan mandates that at least 50% of board seats in these AI companies must be publicly appointed or democratically elected, shifting control away from shareholders. It also requires major tech companies such as Google, Microsoft, and Amazon to divest their controlling interests in frontier AI development to ensure operational independence.

Additionally, the proposal supports the concept of public ownership of AI, referencing Senator Bernie Sanders's American AI Sovereign Wealth Fund Act, which aims to transfer 50% equity from large AI companies into public hands. This reflects a belief that the public, having funded the research and provided the data for AI, should benefit from its existence.

Industry leaders themselves have voiced concerns about the concentration of power in AI development. Anthropic CEO Dario Amodei has expressed discomfort with decisions being made by a select few companies, and AI pioneer Geoffrey Hinton has warned of AI replacing human cognitive abilities, drawing parallels to the Industrial Revolution's impact on labor.