In Summary
- The Allianz Pulse Survey takes the temperature of European households each year on the economic, political and societal issues shaping their lives, with the 2026 edition focusing on artificial intelligence. As AI moves into everyday life and work, the survey explores how people use the technology, what they expect from it and how prepared they feel for the changes it will bring. For its eighth edition, we surveyed a representative sample of 1,000 adults in each of Germany, France, Italy, Spain, Poland and Austria, adding the UK and US for the first time. This provides a transatlantic perspective on how people are experiencing and anticipating the AI transition.
- Europe is embracing AI, but adoption is less advanced and less work related than in the US, with large differences within Europe. Across all eight countries surveyed, 48% of respondents use AI at least weekly, with the US and Spain leading at around 55%, while Germany, France and Austria remain in the low 40s. The generational divide is particularly striking: 77% of Gen Z use AI weekly, compared with just 23% of Boomers. Moreover, adoption has yet to translate fully into workplace use: among AI users, 53% use it mainly for private purposes, 20% mainly for work or study and 27% across both contexts. Just under half therefore use AI at least partly for work or study, with the share highest in the US (52%), the UK (51%) and Spain (49%), but only 37% in Germany. Germany stands out as the country where AI adoption is most concentrated in private life, while Spain shows that high AI adoption and workplace use are not uniquely American.
- Europeans see the benefits of AI for themselves, but are considerably more skeptical about what it means for the economy and jobs, especially compared with Americans. Net sentiment towards AI is -34% overall, meaning negative views exceed positive views by 34pp. It deteriorates to -43% for the economic impact and -67% for jobs. The views about the job impact of AI are deeply negative, even as positive productivity expectations exceed negative expectations by 30% across all respondents, while 35% expect little or no impact. The US provides a clear contrast on employment: its jobs sentiment stands at -44%, still negative but considerably less so than the survey average. Within Europe, the picture varies strongly: Austria (-81%) and France (-75%) are the most pessimistic about the AI impact on jobs, while Spain combines high AI adoption with less negative attitudes towards AI’s expected job impact (-66%). Across countries, young, urban and high-income respondents are the demographic groups most positive about AI. More broadly, the observed AI skepticism is closely associated with an overall weak confidence in economic progress: respondents with negative views of AI also tend to hold gloomier views of both the economy and the future.
- As AI adoption spreads, concerns about its impact on jobs are intensifying, even though most people expect task transformation rather than outright job destruction. Among countries surveyed in both years, net sentiment on AI’s impact on jobs fell by around 12pp, from -55% in 2025 to -67% in 2026. The deterioration was particularly pronounced in Europe: Austria’s jobs balance fell by around 19pps, while Germany and France also recorded double digit declines. Yet this is not primarily a story of expected mass unemployment. Only 7% expect their job to disappear entirely, while 21% expect significant parts to be replaced and 26% expect AI mainly to augment their work. First-hand experience changes expectations substantially: 41% of workplace AI users anticipate augmentation, compared with only 11% of non-users, although they are also more likely to expect task replacement (40% versus 17%). Exposure appears to make both the risks and opportunities more tangible, particularly the potential for AI to support rather than replace workers.
- Europe risks entering the AI transition with too little investment in workers, while training continues to favor those already ahead. Only 17% of all employed respondents across countries have received AI training at work, compared with 30% in the US. The lack of training opportunities is a major bottleneck: 44% report that their employer offers no AI training at all. The European range is wide, with participation falling to just 10% in Poland, while training is also significantly more common among tertiary educated workers. This suggests that access to training risks reinforcing existing divides in skills and opportunity. The transatlantic contrast is particularly concerning: The US combines more intensive AI use with greater employer-led preparation, while parts of Europe combine limited workplace adoption with stronger fears of labor-market disruption.
- People fear that AI will widen inequality and want stronger safeguards against its risks. Across the entire sample, 37% expect AI to increase inequality, 34% think it will stay the same, compared with only 9% who expect a reduction. Austria records the strongest expectation of widening inequality, while Poland is the least concerned, though the balance is positive everywhere. People worry less about what AI might become than about harms it already causes: misinformation (25%), data privacy (21%) and job losses (17%) top the list, ahead of loss of control over advanced AI (16%). The net support for stronger regulation, calculated as the share of respondents in favor minus the share opposed, increased by 17pp, from +61% in 2025 to +78% in 2026. This does not signal outright resistance to AI: Spain combines high adoption with strong support for regulation, showing that enthusiasm for the technology and demand for safeguards can coexist.
- For Europe, the central challenge is therefore to accelerate productive AI adoption while ensuring that its benefits are broadly shared. The survey points to five priorities. First, move AI from private experimentation into productive workplace use by lowering implementation costs, especially for SMEs. Second, close the training gap through employer-led, occupation-specific learning. Third, shift labor-market policy from protecting existing jobs to supporting transitions into new tasks and roles. Fourth, make sure the productivity dividend is broadly shared through skills, wider household participation in capital markets and tax systems that adapt to shifts in income from labor to capital. Fifth, strengthen trust through targeted, proportionate regulation addressing misinformation, data misuse and labor-market risks without stifling innovation. Ultimately, success will be measured by AI's ability to deliver broadly shared productivity gains, stronger employability and greater confidence in economic progress.