Bank of America claims AI productivity could grow 10x over next decade, but current economic data shows only 0.1% gain

BofA's research team published a report arguing that AI's eventual productivity impact could be 10 times larger than current economic output, but the bank acknowledged that measured AI productivity gains are currently only 0.1% per year — small relative to global growth of 3.5%.

The gap reflects that while AI delivers task-level gains (software developers completing 55% more work, customer support agents resolving 14% more tickets, professional writers finishing projects 37% to 40% faster), only about 20% of workplace tasks are transformable by AI and only 23% of those are currently cost-effective to automate at today's prices.

BofA's 10x case rests on the premise that as AI models improve and inference costs fall — halving roughly every three months — the share of tasks both transformable and economically viable to automate will expand, with each incremental expansion compounding non-linearly; the bank cites work by economist Philippe Aghion published in 2024.

Tyler Cowen forecast AI's contribution to U.S. growth at 2% to 2.5%, citing institutional drag from sectors including government, higher education, healthcare, and nonprofits that will be 'very slow to adjust.'