Time to Reset

AI’s value is in the spread

Neil D. Lawrence

Time to Reset, Financial Times, Bracken House, London

Not a race …

The AI race

  • Characterising AI as a race implies one winner
  • That narrative pulls us toward other people’s problems
  • The UK’s task is diffusion, not a bronze in someone else’s sprint

An old warning

An Old Warning

Value in the spread …

Where the value is

  • The prize is not the cleverest model
  • It is millions of firms and workers building on top of it
  • A country that hosts a frontier lab but cannot diffuse captures almost none of it

Why the demand side …

AI cannot replace atomic human

Inflation of Human Capital

  • Automating mental labour is like discovering the philosopher’s stone
  • Efficiency rises; measured skills inflate and lose scarcity value
  • A supply-side reflex chases what the machine can already quantify

What Does Not Inflate

  • What the machine can quantify, it can automate
  • What matters most is hardest to measure: judgment, care, trust
  • That work lives on the demand side, next to real problems

The missing half

Supply-side policy, demand-side prize

  • We have supply-side policy. We need a demand-side strategy
  • The ministerial menu is shaped by those who sell infrastructure, models and platforms
  • Foreign capital and foreign software do not become British productivity on their own

Central deployment without engagement

The Horizon Scandal

The Horizon pattern

  • Centrally deploying technology on people, without engaging them
  • The failure is not the software. It is the absence of feedback
  • AI announcements keep repeating the pattern

The alternative: ground up

Building from the Ground Up

  • Greater Cambridge Shared Planning: AI analyzing consultations
  • NHS England East: Resource allocation during COVID-19
  • Cancer diagnosis: Clinician-led AI development

Success Factors

  • Local expertise drives development
  • Solutions address real operational needs
  • Professional judgment enhanced, not replaced

Concentration strangles diffusion

Bottled up at the top

  • If a few gatekeepers set the terms, the technology stays at the top
  • That is not only unfair. It is how the UK fails to get paid for the AI transition
  • The useful stuff comes from small firms, everywhere — if they can reach a customer

Sovereignty means having alternatives

Sovereignty

  • Hosting someone else’s infrastructure is not sovereignty
  • A British business stranded by a US boardroom is not sovereign
  • Sovereignty here is options, not a national champion

The instrument we already have

Pro-innovation/growth is not pro-big-tech

  • The DMCCA is targeted, flexible, and can move with the market
  • A pro-innovation agenda is not a pro-big-tech agenda
  • The ask now is pace and ambition, not new legislation

The investment we should worry about

Reframe inward investment

  • Opening the gateways is the pro-investment move
  • The fear has it backwards
  • The investment we should worry about losing is the round that never happens

Campus UK

Universities as Innovation Bridges

  • Neutral conveners bringing together diverse expertise
  • Access to research and human capital
  • Creating spaces for collaboration

Key University Roles

  • Connecting developers with users
  • Facilitating ethical oversight
  • Supporting knowledge transfer

Campus UK

  • Draw a circle around what a train can reach
  • Reward the local paper as well as Nature
  • Civic universities were founded for this

A reset

What a reset looks like

  • Treat the demand side as strategy, not a downstream consequence of supply
  • Use the DMCCA with pace, to keep markets contestable while AI is still forming
  • Measure success by diffusion: firms, public services, and places that can actually use the technology

Thanks!

References

Hayek, F.A., 1945. The use of knowledge in society. American Economic Review 35, 519–530.