The Alan Turing Institute investigated the prevalence of generative artificial intelligence (AI), which creates new material (e.g., text, figures), in the United Kingdom’s public sector. They found that nearly half of respondents were aware of the use of generative AI in their work area and 22% were themselves actively using a generative AI system for work.

The use of generative AI was most common in universities and schools and less common in social care, health care, and emergency services. However, predictive AI (which finds patterns to make predictions) was the most common in health-care services, and perceptive AI (which senses aspects of its environment to support interacting with it) was most common in emergency services. This demonstrates the nuance in how different sectors are engaging with the different types of AI.

Most public sector workers felt positive and trusting about the current and future use of generative AI. In the context of productivity, respondents said generative AI had an especially positive influence on reducing bureaucratic procedures and workloads.

Across the public sector, at least 72% of respondents said AI improved their productivity. One notable concern for most (76%) was a worry about missing out on the benefits of working with AI to improve public services, even though only 32% said guidance on the use of generative AI was clear.

The authors’ work suggests that generative AI is being used across the UK public service sector, even in the absence of clear guidelines. However, the authors warn that this uptake has been disorganized and that there is an urgent need for systematic methods and guidelines to ensure proper use of these services.