A policy briefing has been issued by the Ada Lovelace Institute, on the use of AI Foundation Models (such as ChatGPT or MidJourney) in the public sector.
Foundation models like Chat GPT or MidJourney can produce various outputs and applications, such as text, image or audio generation. The briefing responds to increasing interest in the public sector, including growing informal usage and some formal prototypes and proofs of concept.
The briefing is written against a background of increasing appetite to leverage such tools to enhance public service delivery during growing demand and budget constraints, as well as potential risks, such as biases, privacy breaches, misinformation, security threats, overreliance, workforce harms and unequal access.
The policy Briefing provides recommendations for policymakers and public sector leaders to manage usage through responsible and ethical governance and leadership. It provides guidance, procurement requirements, auditing, public engagement and pilot use cases alongside appropriate training and monitoring guidance to reduce risk. It is delivered alongside a caution that foundation models may be adopted because they are a new technology rather than because they are the most suitable solution to a problem and that public-sector users should, therefore carefully consider the counterfactuals before implementing foundation models.
Importantly, the guidance refers to the Nolan Principles of Public Life, which include accountability and openness.
Click here to read the policy briefing from the Ada Lovelace Institute