According to Charbel-Raphaël, a governance researcher who co-founded CeSIA, most consequential AI policy work happens invisibly within government institutions and international bodies—not in the press coverage or public statements that dominate discourse. While the AI safety community celebrates visible victories like press campaigns and open letters, the real regulatory frameworks get built in ministerial cabinets and behind closed doors. Examples include drafting provisions in the EU AI Act, shaping the United Kingdom's AI research strategy, and organizing briefings that influence executive-level decisions. Raphaël estimates ninety percent of government policy work remains entirely hidden—private memos and conversations targeting specific decision-makers. The core argument: the AI safety community overinvests in visible intellectual work while overlooking insider policy channels that actually build regulations. The implication challenges a comfortable narrative: meaningful influence often requires less visibility and more patience grinding through bureaucratic channels.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton