The AI industry is where banking was in 2006. (We're hiring)

In a new post, CeSIA (Center for AI Safety) draws a stark parallel between today’s AI industry and the banking sector in 2006. Just as Raghuram Rajan warned at Jackson Hole in 2005 about unpredictable risks from complex financial instruments, CeSIA argues AI systems are becoming opaque, risk is migrating to unregulated corners, and incentives reward speed over safety. The post cites the 2008 collapse of Bear Stearns, Lehman Brothers, and others as a warning: structural failures—like self-designed risk models, voluntary supervision, and rating agencies paid by rated firms—were known but uncorrected. CeSIA’s mission over the next 18 months is to build institutional capacity for binding regulations: operational risk thresholds, independent evaluations, statutory disclosure, and international coordination. They initiated and coordinated the Global Call for AI Red Lines, launched at the UN General Assembly by Maria Ressa, signed by 12 Nobel laureates and 11 former heads of state. Their work has been incorporated into the EU's Code of Practice.

CeSIA is now hiring three positions to close what they call a dangerous gap between public opinion and institutional action. A commissioned OpinionWay poll found that only 8% of French people want to accelerate AI development, while 42% want to pause or significantly slow it. The roles: Head of Policy Analysis (lead publications and briefings, manage a team), Head of Communications/Communications Lead (strategy, press, crisis comms; fluent French/English required), and Operations & Executive Associate (support leadership; fluent French/English). Applications close May 22, 2026. Work is based in Paris or remote within the EU/UK. CeSIA is positioning itself as the counterweight to industry self-regulation, aiming to prevent an “AI’s 2008 moment” before it happens.