Politics

Copy article

UK Government's AI Growth Lab: Fast-Tracking AI Adoption Through Regulatory Sandboxes

ended 21. October 2025

UK Government Launches AI Growth Lab: Regulatory Sandboxes to Fast-Track Innovation in Healthcare, Housing, and Professional Services

21 October: Technology Secretary Liz Kendall has announced plans for an AI Growth Lab, regulatory sandboxes where companies can test AI products in real-world conditions with certain rules temporarily relaxed under strict supervision. Initially targeting healthcare, professional services, transport, and advanced manufacturing robotics, the sandboxes aim to accelerate responsible AI deployment while cutting bureaucracy that "can choke innovation."

The announcement comes alongside the Chancellor's Regional Investment Summit, where reforms aim to save businesses nearly £6 billion annually by 2029 by reducing pointless admin tasks. Currently, only 21% of UK firms use AI, but the OECD estimates AI could improve UK productivity by 1.3 percentage points yearly, worth £140 billion equivalent.

Real-world applications could include AI tools helping health workers deliver better patient care on accelerated timelines, reducing NHS waiting lists and time demands on frontline staff. 

In housing, typical development applications currently rack up 4,000 pages of documentation and take 18 months from submission to approval. AI could slash those times, supporting the government's plan to build 1.5 million new homes by end of Parliament.

The sandboxes won't allow regulations to be "switched on or off at will" but will see time-limited restrictions on specific regulatory hurdles under close supervision by tech and regulatory experts. 

Strong safeguards include strict licensing schemes, with breaches stopping testing immediately and opening users to potential fines. Exclusions include consumer protection, safety provisions, fundamental rights, workers' protections and intellectual property rights.

A separate £1 million pot will support the Medicines and Healthcare products Regulatory Agency (MHRA) to pilot AI-assisted tools for drug discovery and clinical trial assessments, keeping all decisions "firmly in human hands." 

The government is launching a public call for views on whether the programme should be run in-house or overseen by regulators themselves.

We want your views:

  • Will regulatory sandboxes genuinely accelerate innovation, or create two-tier systems where only well-resourced companies can participate?
  • How do we ensure "strict supervision" and "strong safeguards" actually protect public interest rather than becoming tick-box exercises?
  • If AI can slash housing approval times, why aren't we questioning why those approvals require 4,000 pages in the first place?
  • What happens when sandbox-tested AI products scale to full deployment - do the relaxed regulations snap back, or does this become permanent regulatory erosion?
  • How will the £1 million MHRA pilot ensure AI-assisted drug discovery maintains rigorous safety standards while speeding approvals?
  • Should worker protections and fundamental rights really be excluded from sandboxes, or does that risk creating blind spots in how AI affects people?
  • Is 21% UK business AI adoption low because of regulatory barriers, or because most AI tools don't deliver the promised value?

5 responses from the Newspage community

Copy all

Star Quote
Copy

Colette MasonVerified
AI Ethics Consultant at Clever Clogs AI
The real question isn't whether we can fast-track AI adoption from 21%, it's whether we're adopting AI that actually makes humans better at their jobs or just creates new problems faster. If housing approvals need 4,000 pages and 18 months, that's a broken process problem, not an AI problem.

Throwing AI at broken systems without fixing the underlying dysfunction just gives you expensive, high-speed chaos with a nice dashboard. Regulatory sandboxes can accelerate innovation, but only if 'strict supervision' means rigorous evidence that these tools improve outcomes for real people, not just processing speed.
Copy

Editor
Scott GallacherVerified
Director at Rowley Turton
Anyone questioning whether AI can truly transform the way we work only needs to look back at previous technological shifts — from the telephone to the computer and the internet. Each revolutionised how we do business, and AI will be no different. The Government’s AI Growth Lab is a positive step towards harnessing that potential, but it’s vital that smaller firms also have access to these opportunities so innovation benefits the whole economy, not just big corporations.
Copy

Editor
Tony RedondoVerified
Founder at Cosmos Currency Exchange
The AI Growth Lab could boost productivity, but success depends on equitable access, robust oversight, and reliable infrastructure with accountability for failures, like the recent AWS outage. Public input on sandbox governance (in-house vs. regulator-led) is vital for transparency. AI isn't a cure-all; the 4,000-page housing approval issue shows bureaucratic inefficiencies. While AI can streamline processes (e.g., automating document analysis), the government must reform redundant regulations and manual systems, not just rely on tech fixes.
Copy

Editor
Mitali DeypurkaysthaVerified
AI Governance Strategist & CMO at AISensa
AI isn’t a problem-solver. It’s a catalyst. It amplifies what’s already there. Think of it as a jetpack: strap it on, and if you’re pointed in the wrong direction, it just helps you fail faster and farther. Before we celebrate new AI sandboxes, we must ask if the underlying processes are even fit for purpose. Why, for instance, does housing planning need 4,000 pages of approval? Without addressing inefficiencies and culture first, AI only accelerates dysfunction. The UK’s low adoption rate isn’t from regulation alone, it’s from misdiagnosis. AI was sold as medicine for inefficiency, when in fact, it magnifies underlying flaws such as weak management, toxic culture or outdated systems. Businesses thriving with AI fix leadership, culture and workflow processes first, removing the 'dead wood' before adding AI to unlock genuine transformation. The future of AI success in Britain depends on this order: get the house in order first, then add the jetpack.

Copy

Scott Taylor-Barr0
Principal Adviser at Barnsdale Financial Management
Sandboxes can be great. New developments can be great. But the question must be asked; who's funding this development?

One of the issues we have seen before with Regulatory Sandboxes is that new entrants are given direct access to regulators and get free or heavily discounted fees.

So, why's that an issue? Because that means that existing market participants, who likely get little or no direct access to the regulator and pay substantial fees to operate, are then footing the bill - and many are not happy about paying for their competitors to develop AI systems and other projects that could potentially be in direct competition with them in the near future.

Many financial advisers, across mortgages, wealth management and protection would be unhappy to see their FCA fees being used to help develop an AI adviser whose very reason for being is to put them out of business, for example.