Peter Conti-Brown, professor at the Wharton School, explains how bank-style supervision and stress testing could help manage the risks of increasingly powerful AI companies.
AI is advancing rapidly — but who is responsible when the risks become too big to ignore?
Peter Conti-Brown, professor at the Wharton School, explores whether the banking system offers a useful model for supervising powerful AI companies.
Rather than relying solely on rules written years in advance, bank supervision involves ongoing conversations between public officials and private institutions to identify and manage emerging risks.
Conti-Brown discusses how AI stress tests could work, what might happen if powerful models are hijacked or misused, and why consumers could ultimately end up holding the bag when things go wrong.