AI industry faces 0.1% productivity gain and widespread uncertainty about deployment, as leaders acknowledge 'nobody knows anything' about how to turn AI capabilities into economic returns

Bank of America estimates AI is currently lifting economy-wide productivity by only 0.1% per year, despite calling AI bigger than electricity and the internet combined. The figure reflects that AI can transform about 20% of workplace tasks, but only 23% of those are cost-effective to automate at today's prices, and automated tasks save roughly 27% in labor costs.

Wharton professor Ethan Mollick told corporate leaders at the New York Public Library that uncertainty pervades the field: 'nobody knows anything. We're all making this up as we go along.' He identified organizational constraints—particularly KPI (key performance indicator) requirements that force conservatism—as the mechanism slowing economic payoff, rather than technological limitations.

AI companies are building their own consulting arms to help clients deploy their models, which Mollick noted is paradoxical: if the models are capable enough to 'destroy all white-collar jobs,' they should theoretically solve deployment questions independently. This signals either the absence of a playbook for current conditions or a classic dynamic where a new player possesses something everyone else needs.