AI adoption in Europe (keynote, SF Fed Economic Research Symposium)
A · Strong evidence A keynote by ECB Executive Board member Philip R. Lane that pulls together the ECB's own large firm, household and bank surveys, a randomised information experiment and model simulations, with sources for every chart; strong official evidence, presented as a speech rather than a paper.
Summary based on the full text·Original: about 30 minutes to read
In short
A speech (30 October 2026) by Philip R. Lane, ECB Executive Board member, at the San Francisco Fed's symposium on AI and the economy. Europe builds few frontier models, but AI adoption and AI-related spending already shape euro-area output, trade and finance. Half of euro-area workers now use AI at work, twice the 2024 share, and firms adopt mainly to improve processes rather than cut labour costs. Firms underestimate how many competitors invest in AI, and correcting that belief raises their own plans. Finance follows tangibility: data centres attract external funding, while training is paid from internal funds.
Key numbers
- 52%euro-area workers who use AI at work in 2026, against 26% in 2024 (ECB Consumer Expectations Survey)
- three hoursmedian weekly time saved by working AI users, about 7.7% of median working time
- 15%AI-using firms citing lower labour costs as a reason, against 62% citing better business processes (SAFE, Q4 2025)
- 0.6 percentage pointsrise in a firm's expected AI investment rate per 1 point rise in the expected share of domestic peers investing
- 12%firms planning AI investment that expect to hire AI specialists, against 46% planning staff training
As reported by the source.
English·Web page·opens ecb.europa.eu
What it reports
- Digital investment is up 73% in real terms since 2014, from 2.3% to 3.3% of GDP, mostly intangible; firms say AI adds about 10% to this year's capital spending plans.
- Firms mentioning AI on earnings calls rose from about 10% to about 40% since late 2022.
- SAFE (Q4 2025): 64% of firms use AI experimentally or moderately, 7% significantly.
- In a field experiment, firms told the true share of peers investing in AI revised beliefs up; domestic peers' plans moved own plans, foreign peers' did not. Average expected AI investment rate: 8.3%.
- Worker time savings are skewed: coding and debugging save six hours a week but only about 15% of workers use AI for it; writing and research are common but save less.
- 41% of workers are uninterested in using AI, including 34% of managers; about half say better training would encourage them, and about half of firms plan AI training.
- A model simulation of a 30% productivity jump in IT services gives 4.4% fewer IT jobs, higher real wages in every group, and 3.8% higher euro-area output by 2030.
- Loans to euro-area AI companies are just under €200 billion; AI exposure is linked to about 5 basis points tighter bond spreads and an equity premium of about 1.3 points a quarter.
How they got it
ECB surveys: the Survey on the Access to Finance of Enterprises (SAFE), the Consumer Expectations Survey, the Corporate Telephone Survey and the bank lending survey, plus earnings-call text analysis, securities holdings statistics, trade data and the MANKIO model. The investment experiment randomly gave half of SAFE firms the true share of peers investing.
Limits
- Time savings are self-reported by workers, and respondents who save more select more tasks, which inflates per-task averages.
- The IT productivity scenario is a model result with an assumed shock, not a forecast.
- A speech: several analyses cite forthcoming or mimeo papers that are not yet public.
Why it matters for companies
For European firms, the useful findings are behavioural: competitors' visible AI investment drives plans more than foreign examples, and most firms underestimate peer adoption. Worker gains are concentrated in technical tasks, so broad chatbot rollouts deliver less than targeted use in coding and data work. The financing gap for intangible spending such as training is real, and boards should treat training as investment, since external finance will not cover it.
Summary based on the full text. Read the original for the full argument and data.Read the original on ecb.europa.eu