Generative AI Reaches the Board Room
Research
19.02.2025
ZEW Representative Survey of CEOs on Generative AI
The majority of CEOs in the information economy and manufacturing industry have already personally used generative artificial intelligence (AI). Their own experiences with generative AI influence how optimistic decision makers are about its potential for productivity. These are the findings of a recent representative survey conducted by ZEW Mannheim among around 1,200 companies in Germany’s information economy and manufacturing industry. It is primarily corporate executives who participate in the survey.
“Around 69 per cent of CEOs in the information economy have already used generative AI tools such as ChatGPT, Copilot or Gemini: 29 per cent of CEOs state that they have used generative AI themselves for both private and business purposes, and 20 per cent each state that they have so far only used such applications either purely for business purposes or for private purposes,” says Dr. Daniel Erdsiek, director of the study and member of ZEW’s “Digital Economy” Research Unit.
CEOs of small companies less experienced with AI
“By contrast, around a third of executive managers in the information economy report that they have not yet tried out any AI language models. However, this proportion is only half as high in large companies with at least 100 employees as in small companies with five to nineteen employees,” adds Erdsiek.
In the manufacturing sector, which is less characterised by office-based activities, almost half of the executive managers report that they have not yet tried out any AI language models. This proportion is also significantly lower for large companies than for small companies.
Productivity expectations increase with experience
Dr. Daniel Erdsiek
“Personal experience with AI language models plays an important role in determining what kinds of expectations CEOs might associate with generative AI. This is also likely to have an impact on the extent to which management pushes for suitable measures to promote the efficient use of generative AI in their own company and thus tap into the productivity potential of this new technology,” explains Erdsiek.
In the information economy, 68 per cent of executive managers believe generative AI will have a positive impact on the productivity of employees who work predominantly with computers. This contrasts with only 13 per cent who expect a negative effect on productivity. CEOs in the manufacturing industry estimate the productivity potential of generative AI at similar rates.
In both sectors, the view of productivity effects is strongly linked to management's own experience in dealing with AI language models. In the information economy, for example, 93 per cent of executive managers expect positive outcomes if they already have personal experience using language models for private and business purposes. Among executives who have not yet gained any experience with generative AI, however, the view is much more pessimistic. Around 21 per cent expect negative outcomes and 37 per cent do not expect any impact on productivity.
Data foundations
These results are part of the “ZEW Business Survey in the Information Economy” from February 2025. Around 1,200 companies from the manufacturing industry and the information economy, which is made up of the ICT sector, media service providers and knowledge-intensive service providers, took part in the quarterly business survey, which has been conducted in Germany since 2011. In each wave of the survey, data is also collected on current key topics in the research field of digital economics. These include current ICT trends, investments in ICT or the diffusion of new ICT applications. The central topic of the current edition is generative AI.
