LessWrong strategist proposes using data center bans to regulate AI models

In a LessWrong post from June 2026, user less_raichu outlines a pragmatic strategy for AI safety advocates: piggyback on the widespread backlash against data centers. Rather than extolling the virtues of AI infrastructure, they argue that aligning with local opposition—even if it means supporting data center moratoriums—can be a lever for regulation. The post suggests slipping clauses into county-level bans that prohibit AI companies from deploying 'untested models,' forcing legal battles that raise public awareness. The author also mentions supporting Alex Bores's congressional campaign, but sees local tactics as more immediate and actionable.

Commenter AnthonyC pushes back, noting that data center opposition is usually driven by local concerns (energy, water, noise), while AI risks are global—so local bans may simply shift capacity elsewhere without meaningful regulation. less_raichu counters that the anti-data-center fervor already includes global fears (job loss, oligarchy), and advocates for a guerrilla regulatory approach: even failed ballot measures can educate the public. The debate highlights a growing split in AI governance circles between top-down federal policy and bottom-up grassroots pressure.