AI adoption at the Met Office was a change management problem, not a technology one. That was the headline from Professor Kirstine Dale, the Met Office’s Chief AI Officer, at aiblLIVE London 2026. A wave of AI weather models in late 2022 put the organisation under pressure. A 60-year-old home of physicists suddenly had to justify its people and its new £1.3 billion supercomputer.
The response was to keep the physics and bring everyone with it, while a few funded teams pushed for breakthrough science. Here is how it worked.
All figures as stated on stage by Professor Kirstine Dale.
A step change in weather models built by others. Around December 2022, new models from large technology companies suddenly looked far stronger, and the questions landed fast.
We’ve just signed the contract for a 1.3 billion pound supercomputer. Folk were like, what do you need a supercomputer for? And why do you need a workforce full of physicists?
Professor Kirstine Dale, Chief AI Officer, Met Office
It ran AI adoption as a two-track change programme. One track raised the floor, so everyone understood what AI was and where it could help. The other raised the ceiling, funding focused teams for breakthrough science, including the Fastnet model built with the Alan Turing Institute.
The floor work became the AI for Everyone programme: training, surgeries, tools, hackathons and seminars, backed by clear policies. The aim was to make AI useful across the organisation, not only in specialist teams.
Because AI models are fast but not self-sufficient. Dale said the AI models run tens of thousands of times faster than the physics-based ones, which is why both belong together.
Physics supplies the training data. It gives the deep process understanding needed to explain a safety-critical forecast. And it covers a changing climate that AI models have not seen in training.
Change starts and ends with people, and AI has to serve the mission rather than become it. The workforce moved from worried, to confident, to curious about where to use AI well.
If I was only allowed one lesson, the lesson would be that change starts and ends with people. AI adoption is really not a technology challenge.
Professor Kirstine Dale, Chief AI Officer, Met Office
Professor Kirstine Dale is Chief AI Officer at the Met Office. She leads the responsible use of AI in weather forecasting, climate science and national resilience. She spoke on the Workforce track at aiblLIVE London on 20 October 2026.
If your AI rollout is stalling because people see it as a threat, this is the model to borrow. The full PDF goes deeper than this page, in Professor Dale’s own words.
For COOs, CHROs and transformation leads rolling out AI where trusted experts and safety-critical decisions are involved. Five pages, about a seven-minute read.
