Introduction
I’m delighted to be right here at RUSI.
A novel organisation
… with a proud historical past of main debates on the safety problems with our age
… and with a robust ambition: to construct a safer, equitable and steady world.
From the revolutionary impression of steam energy in the nineteenth century, to nuclear submarines in the twentieth, and the significance of cloud computing for our present defence … you’ve lengthy argued that know-how is a defining function of nationwide safety.
Which is why I’m so happy to be giving this speech at RUSI.
The age of disruption
I need to begin with 2 truths about the world we face at this time.
First: the geopolitical settlement of the final 40 years has ruptured … and lots of would argue, is gone for good.
And second: know-how is disrupting our economies and societies in methods unimaginable just a few years in the past.
The indisputable fact that these 2 forces ought to come collectively at this second will not be a coincidence. Far from it.
They feed and gasoline each other.
Geopolitical rivalry is driving the tech increase.
Why has China poured billions into the semi-conductor trade? To catch up. Why does the US make investments billions into drones and autonomous warfare? To keep forward.
When rival international locations glimpse a new know-how they rush to construct it first, so that they lay declare to the future.
And the compounding nature of know-how means securing a vanguard in a single growth can inexorably result in one other.
Technology can also be driving political discord.
Through the energy of algorithms, and the attain of social media, tech and AI are propelling the populist politics that form our more and more divided and unstable world.
There’s little level in asking which got here first. The know-how and geopolitics gasoline each other.
And create an age of disruption.
AI: The defining forex of the future
The know-how which drives this disruption greater than some other is Artificial Intelligence.
I studied historical past at college; not laptop science or engineering.
And what historical past teaches us is that the nations which pull forward are those who grasp the defining forex of their age.
In the previous this was the measurement of your navy, the unfold of your railways, the attain of your energy grids. The infrastructure for commerce, commerce and militarization.
Today the defining forex is AI.
And the international locations which harness AI will not solely lead the race to treatment ailments, uncover new supplies and create trillion greenback firms … however additionally construct far extra highly effective militaries.
Put merely, AI is now the engine of financial energy and exhausting energy.
And the future is coming at us quick …not in the subsequent few a long time however the subsequent few years.
In round 7 years, AI fashions have progressed from finishing duties that toddlers can do to surpassing PHD stage intelligence.
And it is just dashing up, with AI mannequin capabilities now doubling each 4 months as an alternative of each 8.
Extrapolate this development and by the finish of subsequent 12 months AI will be capable of full in hours what would presently take software program engineers weeks.
And management of this extremely highly effective know-how is changing into more and more concentrated, with 70 per cent of world AI compute now managed by simply 5 firms – up from round 60% a 12 months in the past.
Faced with the energy and velocity at which this know-how is creating we should ask ourselves… how can we make this know-how work for Britain and ship a greater, safer future for all?
We should form this know-how, not simply be formed by it.
We should guarantee its growth displays our pursuits and values and performs to our strengths.
And we have to be able to seize the alternatives, in addition to construct resilience to the dangers.
In brief, we will need to have larger management.
AI sovereignty
We aren’t the solely nation to be grappling with these points and the query of AI sovereignty.
Each nation has its personal interpretation.
For some it means constructing the whole AI stack – the power, chips, information centres, fashions and functions – to develop into fully self-sufficient.
For others it is about creating the proper establishments, governance and regulatory oversight of AI, to match their nationwide tradition.
This authorities believes AI sovereignty will not be about isolationism or making an attempt to pull up the drawbridge and go it alone.
We will proceed to use the finest know-how and welcome inward funding as a result of that’s what our public providers and financial system demand.
For Britain, AI sovereignty is about lowering over dependencies and growing resilience in key nationwide strategic priorities, as the Prime Minister has rightly argued.
So we safe larger management and larger leverage over the points that matter most.
And in order for you true leverage for your nation, that you must be a keystone in the world tech structure – an indispensable associate.
This requires 2 key shifts in our method.
First, a decisive transfer in direction of backing extra British AI firms, particularly in areas the place now we have actual strengths.
And second, by working extra intently with our worldwide companions, notably different so-called center energy nations, together with on setting the requirements for how AI is deployed.
Everyone wants confidence in how AI instruments will probably be used, wherever they’re in the world.
Safety and safety isn’t to stifle innovation or velocity. It is about future proofing what we do, working with our companions.
Because on this more and more unstable, bi-polar world we consider Britain’s nationwide pursuits are strengthened, not diminished after we work with like-minded nations.
Backing Britain and British AI
We are already making actual progress in direction of these targets.
Earlier this month I launched Sovereign AI – a main step which I consider will probably be one of the single most essential issues this authorities does to construct a greater future for our nation.
Sovereign AI combines the velocity of enterprise backed by the weight of a nation to grab the alternatives now we have proper right here in Britain: our very good expertise pool, world main universities, vibrant tech ecosystem, and pragmatic not dogmatic method to regulation.
Sovereign AI will make investments £500 million in British AI firms to begin up, scale up and win globally … and supply the key to unlocking a lot wider authorities help the place it makes a actual distinction.
That means
… offering absolutely funded entry to the UK’s largest tremendous computer systems, with a number of firms already on board and thirty extra in the pipeline
… quick monitoring world expertise, with tremendous precedence visa selections in a single working day and as much as 10 visas for R&D expertise fully value free
… working seamlessly with the British Business Bank and its £2 billion of funding to seamlessly take firms to the subsequent stage
… and mobilising the energy of authorities procurement to again the better of Britain, together with £400 million ringfenced by the Ministry of Defence explicitly to help British innovation together with innovative applied sciences like AI.
Taken collectively these steps mark a decisive shift in direction of making certain Britain is an AI maker not simply an AI taker.
Sovereign AI has already made 2 direct investments.
… in Callosum: a superb crew constructing the AI infrastructure of the future…
… and simply yesterday we introduced our funding, alongside the British Business Bank, in Ineffable Intelligence … based by David Silver, a world-leading AI scientist and considered one of the key architects of DeepMind.
And at this time I can announce we are going to go additional, with a new AI Hardware Plan which I’ll launch at London Tech Week in June.
The world AI chips market is rising at an annual fee of 30% and anticipated to achieve one trillion {dollars} in the early 2030s.
If Britain may safe simply 5% of this market it could convey fifty billion {dollars} in income to the UK with tens of 1000’s of excessive paid jobs in tech.
There are those that say this race is already misplaced. That it’s too late to problem the dominance of the US or China in AI chips.
But I don’t settle for such defeatism.
We have a wealthy historical past of excelling at {hardware}.
The first programmable laptop. The first digital reminiscence. The first business laptop, first parallel laptop and the first broadly used chip IP mannequin all occurred proper right here in the UK.
ARM’s CPU structure is the most generally used processor design on earth. In virtually each smartphone, pill and in increasingly AI servers throughout the world.
And AI compute is quickly diversifying, with completely different {hardware} wanted for completely different duties. This shift gives actual openings for new entrants and specialist {hardware} that couldn’t have been predicted, even just a few years in the past.
We’re investing £100 million in ARIA’s scaling compute programme together with £50 million for their scaling inference lab, so British startups can take a look at and showcase that their new AI {hardware} works.
Alongside ARM, we have sensible new British firms like Fractile, Olix, Lumai, Optalysys and Salience Labs … and we’re decided to help much more with our new AI {hardware} plan.
So sure, this a aggressive market. But we’re a aggressive nation. And I’d argue we are higher positioned than virtually some other nation to run the race to develop new, faster, smarter and extra power environment friendly {hardware} which is essential for the way forward for AI and our future prosperity.
Working with our allies
So there’s a lot we are already doing to safe our AI future, proper right here in Britain.
But there’s extra we are able to and should do to construct our sovereign capabilities and enhance our leverage by working with our allies, particularly different center energy nations.
We already have a Strategic Science and Technology Partnership with Germany, including a £6 million joint quantum project.
Last 12 months our Prime Minister and President Macron agreed to construct on the Entente Cordiale with a new Entente Technologique.
Our Growth and Innovation Partnership with Canada contains joint work on AI safety.
Our Digital Partnership with Japan covers AI, cyber safety and information.
And we not too long ago invited overseas ministers from international locations together with Australia, Canada and the Republic of Korea to debate the most urgent geo-economic challenges of our age, together with know-how.
I’ve been having additional conversations with lots of my counterparts.
They see the similar challenges and alternatives as we do.
The have to construct stronger sovereign AI functionality.
To form this AI future, not be formed by it.
And guarantee AI works for all, not simply a highly effective, unaccountable few.
I consider there are 2 areas the place we ought to now do extra, to construct complementary strengths and make ourselves an indispensable associate.
The first is collectively investing in elements of the AI worth chain the place we share a bonus and aggressive edge.
Look at our work in one other transformative know-how – Quantum.
This month I hosted the fifth assembly of the Quantum Development Group: a discussion board of 13 like minded international locations working collectively to advance the accountable growth of quantum applied sciences, together with working collectively on funding. This is strictly the type of position we in the UK needs to be enjoying in AI.
And second: combining our efforts to spice up resilience, together with on AI safety.
One of Britain’s genuinely world main strengths is the AI Security Institute (AISI), an organisation I am extraordinarily pleased with.
Earlier this month Anthropic introduced that their new Mythos mannequin may uncover novel cyber vulnerabilities.
AISI examined and verified these claims, and its evaluation was broadly reported and heralded throughout the world.
And different fashions have since proven comparable capabilities, like OpenAI’s GPT 5.5 launched final week.
At the similar time, highly effective open-source AI fashions are quickly following the frontier.
This development has profound implications for our nationwide safety, our nation, and our future.
The UK is in a robust place to take care of these challenges, with each the internationally recognised National Cyber Security Centre (NCSC), and our AI Security Institute (AISI).
Many international locations are enthusiastic about working with us on these essential points.
So in July, at the subsequent assembly of the worldwide community of AI Security Institutes, which the UK Chairs, we are going to publish finest observe on the science of evaluating AI fashions.
Working with our allies to share our analysis and expertise of mannequin analysis, construct capabilities and strengthen our resilience collectively.
None of this needs to be taken as weakening our deep, shut and enduring relationship with the US.
Far from it.
We stay dedicated to the UK US Tech Prosperity Deal – the first of its form – with all its potential advantages.
But, as the Prime Minister has rightly mentioned, we should step up – in NATO and throughout Europe – on defence and nationwide safety, and an important a part of this is know-how.
Because stronger nations make for higher allies.
And that’s what we’re decided to be.
Conclusion: Rebuilding Britain for the new world
So that is our plan.
Backing British AI.
Building accountable tech, together with with our allies.
So we handle the twin pressures of geopolitical and technological disruption …
… and rebuild Britain for the new world.
Indeed I would go additional and argue Britain is uniquely positioned to lead these efforts …
… due to our large strengths in AI, and in AI safety
… and as a result of this authorities believes that an lively and muscular state is essential to seizing the alternatives and managing the dangers AI inevitably brings.
I need to end with a wider level, about the political alternative I consider we now face.
I know there are lots of folks on this nation who’re deeply fearful about the state of the world, together with what AI will imply for their jobs and their youngsters’s futures.
We’ve already heard calls, together with in Parliament, to ‘Pause AI’.
I feel doing so could be a double betrayal.
A betrayal of British expertise and British pursuits.
It would ship a message that Britain is closed to new concepts and new alternatives. That a rustic so wealthy in expertise, innovation and enterprise has put an ‘out of office’ signal on its entrance door.
If we retreat from progress we retreat from the world, leaving this highly effective know-how to be exploited by different nations to their benefit and our drawback.
The alternative isn’t between a world that has AI and one which doesn’t.
It is a alternative between a world the place we form our AI future, based mostly on our personal pursuits and values, or the place we are left at its mercy and whim.
This authorities is decided to guide our nation into a greater future.
We have a new technology of patriotic British founders and traders who need to construct formidable and accountable AI, and cement Britain’s distinctive place in the world.
We have a authorities that understands the state can transfer quick and work in other ways – as now we have seen with the AI Security Institute and Sovereign AI.
[Redacted political content] We are as soon as once more asserting the significance of our worldwide alliances as a result of we know can obtain extra collectively than we can do alone.
We are standing on the cusp of a technological British renaissance.
And these scientists, these founders, and this authorities won’t let it cross us by.
We will seize this second and alternative for Britain.
For our financial prosperity, and our nationwide safety.
To construct a future that works for all.
Thank you.