Deputy Prime Minister speech on AI for Public Good
A speech by Deputy Prime Minister, Oliver Dowden at Imperial College.
INTRODUCTION
Ladies and gentlemen鈥
The story of technological advancement is one of constant evolution鈥
鈥 punctuated by game-changing innovations.
In my lifetime, the personal computer, the internet, the smart phone, have all made the tech world - and our interaction with it - unrecognisable.
And they have all - in turn - transformed the way that citizens interact with government, and with public services.
I believe another such game-changer has arrived鈥β
鈥 in the form of transformative AI models - including Large Language Models - that enable computers and humans to interact in totally new ways.
The last fourteen years has been a period of incremental tech improvements.
The digital interfaces we use are largely recognisable.
Yes - we have seized new opportunities鈥
鈥 such as rolling out gov.uk鈥
鈥 and making our services 鈥渄igital by default鈥.
But many of the systems that we use have not kept up with advances鈥
鈥 indeed some of them, I鈥檓 afraid to say, have not moved on at all.
Modern AI has the potential to fundamentally change the way that public services operate within just a few short years.
Indeed, if we are still working off the same systems - and in the same way - in another 14 years鈥 or even frankly another two or three鈥
鈥hen we will have failed to embrace the opportunity that now lies before us.聽
OPPORTUNITY
And so, just as the UK is leading the world in the field of AI safety鈥
鈥 the Prime Minister has asked me to ensure we are leading the world in the adoption of AI across our public sector.聽
The opportunity here is hard to put a value on鈥β
鈥 although I notice the IPPR have estimated that there is the potential to save 拢24 billion each year from roll-out of these new technologies.聽
So for me it鈥檚 only by the rapid adoption of AI that we will drive the savings needed to put us on a sustainable path to a smaller state and better delivery of services.
The pace of change is such that new opportunities are being uncovered literally on a daily basis, and a new world is opening out before us鈥β
AI is potentially - and I don鈥檛 say this lightly - a 鈥榮ilver bullet鈥欌
鈥 it dangles before us the prospect of increased productivity, vast efficiency savings, and improved services.
We are already beginning to see glimpses of what these tools have to offer鈥β
鈥 and so I鈥檇 like to paint a brief picture of what the world might look like if we get this right:
VISION OF SUCCESS
In healthcare - AI diagnostic tools could transform primary care鈥
鈥ith appointments transcribed in real time by ambient AI, then instantly producing prescriptions and referrals鈥
鈥β scans read by AI with far greater accuracy 鈥
鈥 and medicines tailored to individuals based on their genetics - again using AI.
In education - 鈥 AI could help eliminate excessive paperwork 鈥
鈥reeing-up teacher time to focus on what they do best鈥β
鈥I assistants could help teachers to adapt lessons to the specific needs of each pupil鈥
鈥 and AI-augmented reality can take interactive learning to another level.
In crime prevention - AI can direct police to where they are most needed鈥
鈥 spot patterns of criminality to discover culprits quicker than ever鈥
鈥nd help keep the streets safer for everyone.
And in all kinds of public sector casework - from immigration processing to benefit claims - AI can be used to summarise complex information鈥
鈥 enabling expert case-workers to spend more time actually making decisions.聽
I could go on nearly forever to cover all areas of public administration鈥
鈥 because there are very few areas of the public sector that don鈥檛 have the potential to be enhanced by these tools.聽
HOW DO WE GET THERE?
The question, though, is how do we get there?聽
I believe the measures we are bringing forward put in the structures, resources, and mindset鈥
鈥 to put the UK on the fastest path to successful adoption of public sector AI.
Taking advantage of our unique strengths鈥
鈥 to revolutionise public services for everyone in the months and years ahead.
Last year, I established a small team of data scientists, engineers and machine learning experts at the heart of Government - the Incubator for AI - or 鈥榠.AI鈥 - under the energetic leadership of Dr Laura Gilbert.聽
The idea of these experts was to work with departments to target the biggest opportunities to both save money and deliver better public services.
The quality of applicants for this program has been phenomenal.
It is incredibly exciting to see such talented technical people choosing to enter public service鈥
鈥 bringing in new ideas to help change the way government delivers services.
In a few short months this team of just 30 individuals have instigated 10 pilot programs, including鈥
-
AI to flag fraud and error in pharmacies - that costs the taxpayer 拢1 billion every single year.
-
A tool that will read and summarise responses to Government consultations, this says something about the scale of Government consultations, but this could save up to 拢80 million a year in central government alone鈥
-
And AI algorithms to help move asylum claimants out of hotels more efficiently鈥 helping to save further millions.
And I can also announce our intention to roll out a new gov.uk chatbot that will provide an interactive interface for people to better navigate Government information and services.
But this is clearly just the very start鈥
鈥 want to ensure that - where these pilots have proof of concept - we can scale them up as fast as possible鈥
i.AI scale-up
鈥nd so, I can announce today that we will more than double the size of i.AI - to 70 people - recruiting the very best of British talent to drive this work across the public sector.
This unprecedented influx of cutting-edge expertise into Government will enable us to design, build and - crucially - implement AI swiftly and at scale鈥
Of course, there is still a huge role for the private sector - and I welcome the collaboration that we have with so many of the businesses in this room today.
Nothing will match the strength and depth of the private sector AI innovation that is happening right now - and as all of you know so much of it here in the UK.
But I believe that by embedding experts at the heart of Government鈥
鈥 and upskilling public servants to utilise these tools鈥
鈥e will set ourselves up to deliver the benefits to citizens as quickly - and as efficiently - as possible.
HORIZONTALS
The other reason it is so important to have this team at the centre of Government is to ensure that - as AI rolls-out across the public sector - we adhere to the following principles:
鈥 sharing best practice鈥
鈥eploying individual models to multiple use-cases鈥
鈥 finding economies of scale..
鈥 and, crucially, ensuring interoperability.
Although I don鈥檛 claim for the moment to have the expertise needed to actually build AI models鈥
鈥 I can see that - like so many great inventions - there is something beautifully simple about what they are actually doing.
Indeed, when you boil it down, I think there are four ways AI can be applied to much of public sector activity鈥
鈥 spotting patterns of fraud and error;聽
鈥 helping the public to navigate services;
鈥 managing casework;
鈥 and automating internal processes.聽
And so the i.AI team have been looking across these applications with those principles in mind鈥
鈥 And I have agreed with the Treasury that we will make all funding for Government AI projects contingent on departments collaborating with i.AI.
Never again should we be investing money in IT systems without considering how to make them as efficient and interoperable as possible鈥
鈥 or without robustly challenging both the timelines and the costs to deliver better value.
I want to ensure that where we develop a tool for one department - we are considering where else it could be deployed.
MINISTERIAL FORUM
And do to facilitate this discussion鈥
鈥o ensure departments are fully integrated into this cross-government effort鈥
鈥 we need a regular dialogue between all those involved across government.
And so I am convening a meeting of the National Science and Technology Council on AI for public sector good 鈥
鈥 alongside my Co-Chair, Michelle Donelan - our fantastic Secretary of State for Science and Technology.
Every department has now designated a specific minister to be responsible for AI in their area鈥
鈥 and I have asked for them to meet on a regular basis.
In the Cabinet Office, this work will be led by Minister Burghart鈥
鈥 and I want to thank him for the passion, purpose and drive that he has brought to the programme so far, as is often the case when you run a department you get to stand up and make the announcements, but actually Minister Burghart who has actually done the work to bring Government together to do this.聽
WIDER PUBLIC SECTOR JOIN-UP
Of course, central Government can only take this work so far鈥
To truly maximise the benefits on offer we need to work with bodies and agencies right across the public sector.聽
And so I am delighted to announce today that i.AI will sign a 鈥楥ollaboration Charter鈥 with NHS England.
This first-of-a-kind initiative will provide a framework for our experts in the incubator to support the NHS to identify and deploy AI solutions that improve services for patients.
And I would urge other public sector bodies to consider doing exactly the same thing, I think it can bring enormous benefits
RESOURCING
There is no shortage in the Government鈥檚 ambition to use AI for public good.
We have put the expertise and the structures in place鈥
鈥 and we are making progress on our early pilot projects鈥
鈥ut we also appreciate the investment that will be needed to make good on our ambition to see the UK leading the pack.
And crucially, investment will be required both to improve services and cut costs鈥
But also to pave the way for a leaner public sector.
MITIGATING RISK
Through all of this, we are conscious of the need to guard against the risks that have rightly been flagged.聽
And, while every effort will be made to eliminate bias, misinformation, and hallucinations鈥
鈥 ultimately, we are very clear about the need for human oversight鈥
鈥 and a clear distinction between AI suggestions and support on the one hand鈥
鈥nd human decision making on the other.
CONCLUSION
I believe we can take the worst things about public services鈥
鈥hether that鈥檚 the time-wasting, form-filling, pencil-pushing, computer-says-no, the mind-numbing-ness of it鈥
鈥 and the kinds of things that make us want to tear our hair out鈥
We can take those things and we can turn them around with the help of AI.
This is not about replacing real people with robots鈥
鈥t is about removing spirit-sapping, time-wasting admin and bureaucracy鈥
鈥reeing public servants to do the important work that they do best鈥
鈥 and saving taxpayers billions of pounds in the process.聽
We鈥檝e got the political will. We鈥檝e got the world-class civil service. We have the big data. We have the tech companies.
We are ready.
So let鈥檚 not wait.
聽Let鈥檚 lead the way鈥
鈥nd join me in the AI revolution today.