The speech made by Michelle Donelan, the Secretary of State for Science, Innovation and Technology, on 24 October 2023.
Firstly let me say a massive thank you to Onward.
I said when I first launched this new and exciting Department that people could expect to see a constant drumbeat of action from my officials, from my ministers and my entire team…
What I didn’t expect was for Onward to take that challenge on too!
From your brilliant report on generative AI earlier this year, to Allan Nixon’s Wired for Success Report which gave us insights into DSIT that illumined Whitehall.
And I hear through the grapevine that Onward are publishing another AI-focused report in the coming weeks, so I look forward to reading that.
It is safe to say that the world is now wide awake to the opportunities and the challenges posed by artificial intelligence.
In the last 3 years alone, MPs have mentioned Artificial Intelligence more times in the House of Commons than they did in the entire 30 years before that.
And when I stood up to give my first speech as the Secretary of State for the Department for Science, Innovation and Technology, I made it clear that we were going to be different and we were going to do things differently.
So, when it comes to AI, I think it is especially important because we cannot afford for DSIT to be a normal ‘business as usual’ Whitehall department.
Nor can we deliver extraordinary things without more extraordinary people in our Departmental team.
We need to be as agile, innovative and as accessible just like the entrepreneurs and businesses and researchers that we represent.
And I am pleased to report that we have done exactly that.
Over a matter of months we have added world-renowned AI experts to our Department and Taskforce…
From renowned AI professors like David Krueger and Yarin Gal…
To one of the Godfathers of AI, Yoshua Bengio….
With our skills, Frontier AI Taskforce and our global AI Safety Summit, it is clear that the UK is perfectly placed to lead the international charge on seizing the opportunities of AI whilst gripping the risks.
However, today I want to get beyond the statistics and go deeper into the philosophy that is driving our approach to AI.
Many of you will have heard me talk about AI safety being the UK’s priority, and how we can only truly utilise the extraordinary benefits AI has to offer once we have tackled some of the safety challenges associated with the frontier.
To some this may sound overcautious, or that we are being driven by fear of the risks rather than optimism about the opportunities. But actually I think it is rather about the opportunities that we are focusing on.
Today, I want to smash these myths head on.
Here at Onward, I want to set out how the UK’s approach to safety and security in AI will make it the best place in the world for new AI companies to not only grow but locate here.
Safety and security are key to unlocking innovation.
The country which tackles key AI safety risks first will be the first to fully take advantage of the huge potential that AI has to provide.
That is why the UK is putting more investment into these questions right now than any other country in the world.
Questions like “how do we prevent misuse of Frontier AI by malicious actors?”
And “how do we ensure we don’t lose human control and oversight of this new technology?”
And how can we protect our society including our democratic system.
Think of how air travel – once considered a dangerous new technology by many – is now one of the safest and most beneficial technologies in human history.
We got there by working with countries across the world to make sure we have the right safety mechanisms in place – and now we all reap the benefits of flying safely.
Safety is absolutely critical to unlocking adoption across our economy.
Boosting consumer confidence is what will make the difference between a country taking a few years to adopt new technology into their lives, or a few decades.
I want to make sure the UK is at the forefront of reaping the benefits of this transformative technology.
Our approach to AI has been commended for being agile, open and innovative. But we need more research to guide our approach.
In many cases, we simply don’t understand the risks in enough detail or certainty right now because this an emerging technology that is developing quicker than any other technology in human history.
It took mankind just over a lifetime to go from the horse and cart to the space race.
Yet in the last four years large language models have gone from barely being able to write a coherent sentence to now being able to pass the bar exam and medical exams and who knows what large language models have are set to hold.
So the pace of development is fast and unpredictable which means our focus needs to be on understanding the risks.
And in life I do think its important to understand the problem before rushing to solutions. And with AI this should be no different.
That is why we established the Frontier AI Taskforce and have appointed a world-leading research team to turbo-charge our understanding of frontier AI with expert insights.
The Taskforce is already making rapid progress, forging partnerships with industry and developing innovative approaches to addressing the risks of AI and harnessing its benefits.
The Global AI Safety Summit is also an opportunity to build that understanding, share learnings and establish a network globally to work together to ensure our research can keep up with this transformative emerging technology.
Indeed, one of the key objectives of the Summit is to form an agreement on what the key risks actually are.
By bringing countries, leading tech organisations, academia and civil society together, the UK will lead the international conversation on frontier AI.
The global nature of AI means that international concerted action is absolutely critical. AI doesn’t stop at geographical boundaries and nor should our approach.
But of course, we do also need to make sure we have the right domestic approach in place to drive safe, responsible AI innovation.
Earlier this year we set out a principles-based approach through the AI Regulation White Paper.
Our approach is agile, targeted and sector-specific.
We here in the UK understand that AI use cases are drastically different across different sectors.
A one size fits all system that treats agri-tech the same way as military drones because they both use AI is unreasonable and undeliverable.
By empowering existing regulators to regulate AI in their own sectors where they have their own expertise, we have created the most tailored and responsive regulatory regime anywhere in the world.
We have also supported different sectors with a Central risk function – horizon scanning.
Later this year, we will publish a full response to our White Paper – showing how our approach is keeping pace with this transformative technology.
So, what does this all mean for small businesses?
The regulatory approach set out in the White Paper is specifically designed to be flexible, support innovation and ensure that small, new and challenger AI companies can grow and succeed here in the UK.
And indeed, we are already taking steps to deliver on those aims by working with the Digital Regulation Cooperation Forum to pilot a new advisory service for AI and digital innovators so companies can bring their products more quickly and safely to market.
We want the UK to remain one of the most nimble and innovative places in the world for AI companies to grow.
That is why it is right to reaffirm our commitment to our principles-based approach to regulation whilst also taking bold steps to address risks at the frontier,
investing in world-class research capabilities and working closely with industry and civil society to make sure our AI governance approach keeps pace.
Our proportionate and targeted approach will enable us to foster innovation and encourage companies to grow to catch up with the frontier – because we are not pulling up the drawbridge –in fact what we want to do is give consumers and the public confidence to boost adoption and it will ensure we can seize the opportunities safely.
Far from a race to the bottom, the key AI developers across the world and here in the UK are telling me they are looking for countries where they will have certainty, clarity and support when it comes to building and deploying AI safely.
They want a mature, considered and agile approach to AI that maximises the potential for innovation by mitigating the risks.
They want to open up in countries where consumers are open-minded and excited about using their AI tools to improve their lives, which is why with the global AI Safety Summit, we are not only talking about risks, but also talking about opportunities for the benefit of mankind.
And the proof is in the pudding Open AI, Anthropic have opened their international offices here.
I want the UK people to use AI with the same confidence and lack of fear as we do when we book an airplane ticket.
And I want AI companies to know that the UK is the most up-to-date, agile and economically rewarding place in the world to build their business in.
So, to all the smaller AI companies out there, let me send this message out to you today: the UK is and will remain the most agile and innovative place for you to develop your business.
Safety at the frontier means prosperity across the sector.
We will grip the risks so that we can seize the opportunities.