whiteheat.substack.com/p/getting-th...

If we want political benefits to help Labour beat Reform in 2029, AIGL should prioritise working with small district general hospitals in Britain's left behind towns, instead of business as usual where London gets AI before some parts of the country graduate beyond paper notes.

Professional regulation can be a bottleneck to AI use and may limit productivity benefits from innovation, so it is vital that AI Growth Labs works with professional regulators too.

Electronic patient records mean that monitoring and evaluation could be radically improved by AI agents. DHSC should aim to build a "strategic brain", with a continuous real-time census of key measures of healthcare quality, to inform better decision making.

AI-assisted cognitive behavioural therapy (CBT) and physiotherapy should be priorities to tackle economic inactivity due to long-term sickness, scale support for smoking cessation, and give physios more time to focus on getting patients out of hospital.

In healthcare, AI-assisted diagnosis will be helpful, but most healthcare consumption is by elderly patients who have chronic conditions which are already diagnosed - as a hospital doctor I contribute to a new diagnosis maybe once a month.

AIGL should focus on safety-critical sectors with intentionally redundant regulation, like nuclear, pharmaceuticals manufacturing and healthcare, where process regulation will more quickly be outpaced by AI and AI-assisted innovation.

It was great to have the opportunity to work on this project with and learn from Helen Ewles, @ersatzben.bsky.social, @pedroserodio.com and @freddieposer.com.

We’re actively seeking feedback - if you have any recommendations for improvement, please get in touch!

We hope this becomes a vital guide for policymakers and funders like @ukri.org and @innovateuk.bsky.social.

Link to launch piece: britishprogress.org/articles/wev...

Link to toolkit: rd-toolkit.britishprogress.org

We offer clear, actionable recommendations, identifying the most appropriate use cases for each R&D policy tool.

We consider factors such as suitability for regional and sectoral targeting, market characteristics, technological maturity and implementation complexity.

Some thoughts on political narrative for Labour:
findingstarmerism.substack.com/p/britain-is...

whiteheat.substack.com/p/ai-can-fix...

#metascience #openscience