US Senator Urges Democratic Leaders to Prioritize AI Safety and Consider Banning RSI
On Nov 20, US Senator John Hickenlooper of Colorado wrote to Senate Democratic leaders Chuck Schumer and Chris Van Hollen, placing AI safety alongside affordability and accountability as three top priorities for the Democratic caucus—one of the first times AI safety has been written into caucus-level agenda-setting at this level.
Confirmed
- The letter explicitly states that the next majority leader should "consider banning recursive self-improvement (RSI)" to prevent AI systems from evolving autonomously beyond human control
- Other regulatory proposals include supporting a third-party evaluation ecosystem, addressing cyber and biological risks, establishing accountability rules, building a system to track AI incidents and vulnerabilities in real time, and giving regulators stronger enforcement powers to hold AI products accountable for harm
- AI safety researcher SOhEigeartaigh noted that several US politicians have recently spoken about restricting or banning RSI and has begun collecting such statements; this letter is the latest example
Unconfirmed
- According to @MilesBrundage, 7–10 lawmakers already support AI guardrails; this figure is secondhand and the specific list and criteria have not been confirmed by primary sources
Why it matters
- "Banning recursive self-improvement" was previously discussed mainly among AI safety researchers and some frontier labs' safety plans; an incumbent senator now formally putting it into a caucus-level policy letter marks the issue's move from academic agenda to mainstream legislative debate
- It comes as Democrats reassess caucus strategy after winning Senate seats in Kansas, Mississippi and Texas—states they had not won in decades—so ranking AI safety among three top priorities could affect legislative sequencing in the next Congress
2026-11-20 ~ 2026-11-20 · 5 related posts
Primary sources
- Sen. Hickenlooper makes AI safety a top Democratic priority, calls for banning RSI — peterwildeford ·
- Sen. Hickenlooper proposes considering a ban on recursive self-improvement in AI legislation — peterwildeford ·
- US Senator's open letter calls for banning recursive self-improvement in AI — S_OhEigeartaigh ·
- [source] US Senator's open letter calls for banning recursive self-improvement in AI — S_OhEigeartaigh · 2026-11-20
- Senator Hickenlooper urges Schumer to consider banning recursive self-improvement in AI — AndyMasley · 2026-11-20
- [source] Sen. Hickenlooper makes AI safety a top Democratic priority, calls for banning RSI — peterwildeford · 2026-11-20
- Sen. Hickenlooper urges banning recursive self-improvement; 7-10 lawmakers back AI guardrails — Miles_Brundage · 2026-11-20
1 near-duplicate retellings: peterwildeford