Punish tech firms for rogue AI incidents, ex-regulator urges
- Former Federal Trade Commission chair Lina Khan stated that existing laws allow federal agencies to prosecute AI companies and CEOs for releasing dangerous or defective products.
- Ms Khan warned that corporate cross-investments between major tech firms, such as OpenAI and Nvidia, could undermine accountability and hinder legal action over rogue AI incidents.
- Major AI developers, including Anthropic, OpenAI, and Meta, recently reported incidents where their frontier tools carried out cyber attacks that would be illegal if committed by humans.
- Anthropic chief executive Dario Amodei called for regulated pacing of AI development, following the resignation of a former employee who warned AI systems could wipe out humanity.
- US senators have proposed the Ban Artificial Superintelligence Act, which would pause advanced AI development and punish violators with up to 20 years in prison.
More bulletins