From MIT Technology Review · · 1m 40s
Nancy Pelosi’s replacement ready to regulate AI
California senator Scott Wiener wins Pelosi seat and vows to regulate AI before 2026.
In brief
California senator Scott Wiener wins Pelosi seat and vows to regulate AI before 2026. Scott Wiener, who won Nancy Pelosi’s San Francisco congressional seat, says Congress must act on AI risks before 2026 to prevent catastrophic self-improvement. Originally reported by MIT Technology Review.
Scott Wiener wins Pelosi seat
California senator Scott Wiener won Nancy Pelosi’s open San Francisco congressional seat in a tight race.
Wiener focuses on AI risks
Wiener says AI could spin out of control due to recursive self-improvement, a phenomenon where AI advances rapidly without human oversight.
Wiener cites AI report
Wiener called a nonprofit report showing Anthropic’s potential recursive self-improvement 'very concerning' and warned we are running out of time to regulate AI.
Wiener’s AI legislation
Wiener authored SB 1047 in 2024, which was vetoed by California governor Gavin Newsom, and later sponsored SB 53, which Newsom signed.
Public First Action support
Wiener received over $3 million from Public First Action, a nonprofit that receives funding from Anthropic, which opponents criticized as too close to AI industry.
Wiener’s policy ideas
Wiener wants Congress to create a strong federal AI regulator, implement whistleblower protections, and require testing and auditing of AI systems.
State vs federal AI laws
Wiener opposes federal preemption of state AI laws, arguing states are 'laboratories' where nimble regulation can occur.
Wiener’s cautious optimism
Wiener says Congress has a poor track record on tech regulation and hopes for a small disaster to spur action on AI risks.
Wiener’s AI risk warning
Wiener warns that AI could lead to total loss of control if recursive self-improvement occurs, a scenario he calls 'potentially a recipe for total loss of control'.
Here's the gist
Scott Wiener, who won Nancy Pelosi’s San Francisco congressional seat, says Congress must act on AI risks before 2026 to prevent catastrophic self-improvement.
