Question to the Department for Science, Innovation & Technology:
To ask the Secretary of State for Science, Innovation and Technology, what mechanisms are in place to ensure (a) reproducibility and (b) accountability when AI is used in cross-government digital services.
The AI Playbook discusses accountability requirements for the use of AI in the public sector, emphasising human validation for high-risk decisions and user-reporting channels for oversight. Transparency is mandated through the Algorithmic Transparency Recording Standard (ATRS), requiring public disclosure of algorithms in decision-making, with contestability mechanisms for appeals and redress.
The Playbook and Magenta Book stress considering reproducibility and scalability from project inception, supported by rigorous evaluation and quality assurance. The AI Community fosters collaboration across government to prevent duplicated efforts, while the AI Knowledge Hub’ collection of AI use cases supports reproducibility and scalability of projects through shared learning and best practices.