Ofgem, Great Britain's energy regulator, published its decision on 3 June 2026 to establish an AI technical sandbox as a 12-month pilot. Selected participants will test defined AI use cases in a controlled environment separate from live operational systems, using agreed data and evaluation criteria, so no real consumers or energy services are affected.
The decision followed a call for input (July–September 2025) and a consultation that ran from 26 January to 20 March 2026 and drew 32 responses from licensees, operators of essential services, technology providers, system bodies and academics. Respondents backed a phased, cohort-based design, flagged data access (including synthetic data) as the main constraint, and asked for proportionate safeguards, stricter for critical national infrastructure.
Ofgem states that participation is not regulatory approval and that outcomes inform rather than determine later decisions. The pilot is targeted to start in late autumn 2026, subject to published participation guidance and governance, with a lessons-learned summary promised by the end of 2027.
Evidence caveat: as of this entry the pilot has not started, so there are no test results; the practice is rated on design and transparency commitments, not outcomes.
Read the full analysis: https://www.ofgem.gov.uk/consultation/ai-technical-sandbox
Where this practice's information was retrieved from, and when.