ECB: a majority of euro-area workers now use AI at work, saving three hours a week

The European Central Bank published new figures on 26 August 2026 showing that AI use at work in the euro area has doubled in two years. Drawing on the ECB Consumer Expectations Survey, which polls around 20,000 respondents across 11 euro-area countries, authors António Dias da Silva, Laura Lebastard and David Sondermann report that the share of workers using AI in their job rose from 26% in 2024 to 41% in 2025 and 52% in 2026. Users deploy it roughly three days a week.

The median user saves three hours a week, equivalent to 7.7% of median working time, though the distribution is heavily skewed. Accounting for non-users, the ECB puts the economy-wide efficiency gain at 3.8%. Adoption is uneven: 61% among workers with a university degree against 37% among those with lower education levels, and younger workers are around 20 percentage points more likely to use AI than older colleagues.

Why this matters

If you work independently or remotely in the euro area, the useful number here is not 52% but three hours. That is the median weekly time saving among people already using these tools, measured across a large representative survey rather than claimed by a vendor. For a freelancer billing by the hour it is capacity you can reprice or reclaim; for anyone negotiating workload it is a figure with an institutional source behind it. The education and age gaps are worth reading as a warning rather than a description – the gap is in who has been shown how, not in who could benefit.

Context

The ECB’s interest is in whether AI adoption shows up as productivity growth, which is why the blog quantifies time saved rather than tasks automated. The figures sit alongside Eurofound’s European Working Conditions Survey 2024, which found new technologies reshaping European work rather than destroying jobs at the scale often predicted.

What to watch

One caveat to carry: the ECB analysis does not examine remote, hybrid or office working at all, so it says nothing directly about whether AI adoption differs by where people work. Anyone drawing that connection is adding it. The next Consumer Expectations Survey wave will show whether adoption keeps climbing at the same rate or begins to level off.