I’m sure you all know that the government’s new AI Opportunities Action Plan was released this week. I think the plan is a fantastic start. I want to share some of my thoughts on why, and some initial thoughts on what I think it means for us and education as whole.
For those that haven’t read the detail, there are two documents. The first, the action plan, led by Matt Clifford, makes 50 recommendations, and the second, a response document from the government, accepts the recommendations and notes action.
The recommendations are broken down into three sections – I’m going to look at them in a slightly different order to report:
- Secure our future with homegrown AI
- Laying the foundations
- Change Lives by embracing AI
Secure our future with homegrown AI
I’m going to start with this one, as this is the most ambitious, and in many ways most exciting.
The headline slogan is for the UK to “Become an AI maker not taker,” noting “as the technology becomes more powerful, we should be the best state partner to those building frontier AI”
This is quite a shift in ambition, and a welcome one. The UK will not only benefit economically if we get this right but it also puts us much more in control of our own destiny. I wrote about Sovereign AI and why I think it matters for education last summer, and it’s fantastic to see it now firmly in the Governments plans.
How’s this going to be moved forward?
- Recommendation 50. Create a new unit, UK Sovereign AI, with the power to partner with the private sector to deliver the clear mandate of maximising the UK’s stake in frontier AI.
The government response says more details will be shared in the Spring, so I look forward to seeing that.
Laying the foundations
So what are the foundations? A mix of infrastructure, data, skills and talent, and appropriate regulation.
Pleasingly, the introduction starts by acknowledging the contributions of our universities, noting we have:
“Strong fundamental AI research, and high-quality research and engineering talent coming out of our universities, which are some of the best in the world for AI.”
Infrastructure
On the infrastructure side, it’s great to see that power has been included. Questions about the environmental impact of AI are asked at pretty much every event I speak at, and any AI infrastructure plan has to include considerations for this. The report notes:
“Clean and renewable energy solutions are needed to power the increasing energy demands of AI. To identify potential solutions, the Science and Technology Secretary of State and the Energy Secretary will co-chair a new AI Energy Council formed of industry leaders from the energy and AI sectors.”
It goes without saying that access to compute is vital for AI research in the UK, and seeing a firm commitment to this is welcome. I’ll leave others in the research space to comment on whether the numbers work. The aim is for a 20 fold increase AI research computing capacity in the UK by 2030 – starting within 6 months. It certainly seems significantly above what would happen (Moore’s law etc) without intervention.
Jisc, is of course, a key partner in research infrastructure on the UK. The Janet network is a foundational part, as our trust and identity services, cyber security services and licensing services and much more, and we’re looking forward to playing our part in this.
Data
I want to then note some of the recommendation around data, which include the creation of a national data library and three other specific recommendations.
- Recommendation 7. Rapidly identify at least 5 high-impact public datasets it will seek to make available to AI researchers and innovators.
- Recommendation 8. Strategically shape what data is collected, rather than just making data available that already exists.
- Recommendation 9. Develop and publish guidelines and best practices for releasing open government datasets which can be used for AI, including on the development of effective data structures and data dissemination methods.
This is exciting! It’s perhaps natural to jump straight to LLM training data here, and copyright discussions about this rages on as I write this. But that’s probably not the prize here. High quality data to train specialist models (not LLMs) is perhaps more important at the moment for us, as well as data to augment LLM responses.
Education should be part of this – great data will enable great AI solutions. We have already seen this in action with DfE’s hackathon identifying the value of high-quality curriculum mapped data’s impact on the use of AI in assessment and marking, resulting in a £3m investment to make this data and information available to developers.
We need to work out how to make data produced by education part of this. New datasets could help researchers and innovators develop AI tools to enhance curriculum design and student outcomes beyond anything we are currently seeing. If we could join up lifelong learning data and perhaps bring in other factors such as regional and national needs and opportunities, the potential is enormous.
Talent and skills
There is, of course, a focus on attracting and maintaining the highest calibre AI talent in the UK, and this will be welcome by our universities.
I also welcome the fact that the action plan notes that further education and apprenticeships are key to widening access to careers in AI, and that lifelong learning is going to be key for the UK. Our work with colleges in AI literacy means the sector is well prepared for this. Two recommendations look at this:
- Recommendation 17.Expand education pathways into AI.
- Recommendation 19. Ensure its lifelong skills programme is ready for AI.
Change lives by embracing AI
A phrasing and concept that I really like – focusing on the value of AI on improving the lives of everyone. It’s not a silver bullet – we all know that, but it’s got a part to play:
“AI should become core to how we think about delivering services, transforming citizens’ experiences, and improving productivity.”
This, of course, includes education, and it’s great that the use of AI in education was cited as one of the success stories.
“Using AI assistants to do repetitive tasks better and faster, freeing up to 20% of an employee’s time. For example, it is helping some teachers cut down the 15+ hours a week they spend on lesson planning and marking in pilots.”
The suggested approach is described as:
“Scan > Pilot > Scale”
All those that have been involved in our pilots will know this very much aligns with our approach, and we’ve seen that with the right tools, rapid scaling is possible. You only have to look at how quickly use of TeacherMatic spread across the FE sector.
We’ll continue to build on our AI scanning and piloting processes. Our focus this year is on scaling success stories across the sector, starting with AI assistants, based for example on our Learnwise pilot.
These are the most relevant recommendation in this area:
- Recommendation 32.A cross government, technical horizon scanning and market intelligence capability who understands AI capabilities and use-cases as they evolve to work closely with the mission leads and maximise the expertise of both.
- Recommendation 33.Two-way partnerships with AI vendors and startups to anticipate future AI developments and signal public sector demand.
- Recommendation 45. Publish best-practice guidance, results, case-studies and open-source solutions through a single “AI Knowledge Hub”
- Recommendation 49. Drive AI adoption across the whole country.
What next?
I wanted to share initial thoughts. There will be many more discussions about the detail, and how we as a sector can both enable from it and benefit from it. But at this point I really just wanted to share how positive I think this is, and that now is the time for us as a sector to fully engage, contribute, shape, and, where needed and appropriate, lead.
Find out more by visiting our Artificial Intelligence page to view publications and resources, join us for events and discover what AI has to offer through our range of interactive online demos.
For regular updates from the team sign up to our mailing list.
Get in touch with the team directly at AI@jisc.ac.uk