The UK government has announced its AI Opportunities Action Plan, accepting 50 recommendations by the entrepreneur Matt Clifford on how AI could support more efficiency and lead to new market growth. After the initial furor on the launch, what does the plan tell us about what the future might look like, and where are there […]
The UK government has announced its AI Opportunities Action Plan, accepting 50 recommendations by the entrepreneur Matt Clifford on how AI could support more efficiency and lead to new market growth. After the initial furor on the launch, what does the plan tell us about what the future might look like, and where are there still gaps that have to be filled?
What the plan gets right
Let’s start by looking at what this plan gets right. The most important element is that this plan is based on understanding what AI can be good for – automating work and removing repetitive tasks. This key value proposition is that the UK public sector can spend less time doing admin and more time delivering the services working people rely on.
This focus on productivity also shows that the Government gets the bigger problem here, recognising that the infrastructure and associated utilities that are needed to deliver services using AI will be a challenge in the future.
This element is also useful as a way to point at where there is a longer term goal for generative AI. Rather than solely being about generating new content – be that words, images or videos in response to prompts – generative AI can help in automation around back-office tasks that would previously be harder to push through due to the sheer volume of requests.
Public sector organisations suffer because there is no ability to ever say that the general body of work is “done” – instead, there will always be more requests for help, more people to treat, or more services to provide. Reducing admin tasks that are necessary but time-consuming should free people to work on what citizens in the local community will value.
This focus on the public sector, where there is the potential for automation, is also where there is the most opportunity to make real and tangible difference in the short and long term.
What’s possible and practical?
Like any policy document, there must be a degree of positioning the document for political reasons. I doubt that any of us would object to seeing the UK take the lead in how AI is developed and expanded when there are real consequences like saving costs or supporting new businesses to be successful. However, becoming a ‘global IT superpower’ requires more than concentrating on AI and infrastructure.
There are longstanding challenges around talent, funding and education that would need to be addressed too. Getting a workforce in place that can build AI companies is one element; alongside this, there needs to be more insight into how to use AI effectively so that it can deliver in multiple ways.
This maturity curve not only has to look at what is technically possible, but also at the organisational, social and ethical effects that AI will lead to. This is not just about the technology, but also the consequences and opportunities that can be delivered over time. Every organisation – from public sector bodies through to enterprises – will have to go through some form of AI maturity model as they test their approaches, adopt where it is right, and then expand their projects over time.
The next practical consideration is that the Department of Science, Innovation and Technology is due to create a new digital centre of government. This will be useful for providing a central point around AI decision making, but who will drive this body forward and will it have the powers and funding necessary to make a difference? How will it be a centre of excellence that can drive real value for the country and for the companies that are built here? Does the government think it has the right level of skills and experience to do this themselves, or will it rely on the same voices from industry that have been responsible for previous technology initiatives?
Getting the right zones for fast development of AI, from the technology and the data centre locations through to engagement with industry and communities in new ways, will be essential to this process and to increase the chances of success. Should dedicated AI Growth Zones be focused on infrastructure?
The ability to open and re-use public data for AI will be key if we want to gain a competitive advantage. This move has the greatest potential for impact across the UK when it comes to getting AI right. There are safety concerns around sharing or releasing data, and those will always need to be balanced against the need to innovate and achieve goals. Ultimately, this data has more potential to make a difference than simply adding more compute resources, as welcome as a new supercomputer would be.
Right now the biggest problem is how to get AI projects into production with a meaningful use case that will actually make a difference.
Differentiation needed
One element that needs to be expanded on is that, even though there were 50 recommendations, there was no differentiation on the different types of AI technology that might be used. Generative AI and agentic AI are very different to predictive AI or machine learning, and they are all used for specific aims. Picking an approach that is not fit for purpose leads to poorer results and potential long-term harm.
Aside from help with simplifying planning for new infrastructure, there’s little to suggest how the government will be helping support AI start ups. These companies are where the most opportunity for growth – and for financial impact – will come from. By understanding the finer details of how the technology functions from a practical perspective, and how the companies themselves operate, UK government organisations can focus more on what they need, and where they can shape the market to support those businesses to be successful too.
For instance, the option to support “high potential visa routes” for top talent to move to the UK will make it easier for “tomorrow’s talent” to learn here, and then contribute to the economy in a meaningful way. Expanding the pool of people that can work in this area improves diversity and broadening pathways to AI will counteract the risks of AI being inherently biased. This is about building better services or products.
The overall goal around AI is that we all want to see better results from what we put in, whether that is in terms of budgets spent, resources and time invested, or just how our data is used. At the same time, we must understand the process of how AI works to create value. Without these two elements working together, we will miss out on the big opportunities that this plan has laid out.
