The Department for Science, Innovation and Technology has launched a Call for Evidence on proposals to set up an ‘AI Growth Lab’. 

According to the Government, the AI Growth Lab will be “a pioneering cross-economy sandbox that would carefully supervise the deployment of AI-enabled products and services that current regulation hinders”. At its heart is the belief – echoed in the AI Opportunities Action Plan (commented upon here) – that regulatory sandboxes are the answer to identifying and removing regulatory barriers to growth. As the Call for Evidence explains, the Lab would “enable businesses and regulators to trial novel AI products and generate real-world evidence of their impact, which could speed up regulatory approval”. 

Regulatory sandboxes are already in effect in some areas (the Call for Evidence mentions, for example, the FCA’s Innovate Sandbox and the ICO’s Data Protection Sandbox). However, one of the motivations behind the AI Growth Lab is that it would represent a move away from siloed work undertaken by individual regulators and instead work across various sectors allowing the testing of those innovations which “cut-across traditional regulatory boundaries”.  

The Lab would focus on specific sectors that are identified by the Government as aligning with its strategic priorities and where there is a recognition that regularity modification may be necessary. Those operating within the sandbox would be able to “rapidly test and pilot temporary regulatory modifications” with the possibility that some modifications would be made permanent through secondary legislation. However, the Call for Evidence is clear that modifications would be subject to appropriate oversight, and that legislation would specify so-called ‘red lines’ beyond which no regulations could be modified. The Call for Evidence seeks views on what these red lines would be, but the Government intends for them to include matters such as consumer protections, safety provisions, fundamental rights, workers’ protections, and intellectual property rights.  

The Call for Evidence also seeks views on the operating model for the Lab, offering two options. First, it imagines a single, government-operated Lab which would be a central ‘entry point’, responsible for designing and supervising pilots across sectors. Alternatively, there would a series of regulator-operated labs and a lead regulator appointed for each sandbox. A regulatory consortium would then be formed for cross-sector AI applications. 

The Call for Evidence closes on 2 January 2026 and can be read in full here.