US Senator Bernie Sanders and Representative Greg Casar introduced the Ban Artificial Superintelligence Act this week, aiming to permanently ban artificial superintelligence and temporarily pause advanced AI development in the United States, according to medianama.com. The legislation defines artificial superintelligence as AI systems that match or exceed human cognitive abilities across a broad range of tasks and could potentially undermine human control or government authority.
The bill outlines specific dangerous capabilities that would trigger the ban, including AI systems that can plan and execute the disempowerment of humanity, overthrow or undermine the US government, subvert shutdown commands, or conduct unauthorized cyberattacks, medianama.com reported. Sanders emphasized that a superintelligent AI escaping human control would pose a global threat, not just an American or Chinese one. The legislation focuses on the AI system's capabilities rather than specific applications.
This legislative move follows recent incidents involving frontier AI systems that have raised concerns about safety and control, prompting calls for regulatory measures. The bill's approach contrasts with previous AI regulations by targeting systems with broad cognitive abilities and potential to act beyond safeguards. It reflects growing apprehension over advanced AI's risks amid rapid technological progress, aligning with global debates on AI governance and ethics.
The Ban Artificial Superintelligence Act was formally introduced in the US Congress this week, marking a significant step in AI policy. The bill's progress will be closely watched as lawmakers debate the balance between innovation and safety in AI development, with implications for the future of AI regulation in the United States.