UK National AI Strategy 2021–2031
Summary and immediate thoughts — is this a gift to artificial intelligence and technology policy wonks?
This article is an attempt to summarise part of the UK National AI Strategy released the 22nd of September 2022.
Before doing so I thought to clarify what a policy wonk is:
Policy wonk: “a person who takes an enthusiastic or excessive interest in minor details of political policy.”
I am wondering whether this strategy is a gift to artificial intelligence (AI) and technology policy wonks as it has a wide range of threads to follow in the coming year pertaining to AI policy from the UK.
You will see.
The National AI Strategy in the UK was presented to the British Parliament by the Secretary of State for Digital, Culture, Media and Sport. The strategy is a 62-page document readily available online.
It was jointly published by the Office for Artificial Intelligence, Department for Digital, Culture, Media & Sport (DCMS), and Department for Business, Energy & Industrial Strategy (BEIS).
The PR release from the UK government can be found here:
New ten-year plan to make the UK a global AI superpower
Aims to position country as global leader in the governance of AI technologies and includes plans for a white paper on…
Three pillars of the UK National AI Strategy
This national strategy has three pillars.
- Pillar 1: Investing in the long-term needs of the AI ecosystem.
- Pillar 2: Ensuring AI benefits all sectors and regions.
- Pillar 3: Governing AI effectively.
It begins with statements from the secretaries of DCMS and BEIS whereas I find the following statement from DCMS to be of interest:
“This National AI Strategy will signal to the world our intention to build the most pro-innovation regulatory environment in the world; to drive prosperity across the UK and ensure everyone can benefit from AI; and to apply AI to help solve global challenges like climate change.”
— Nadine Dorries, Secretary of State for Digital, Culture, Media and Sport.
The overall headline is also that this is the UK’s ten year plan. As such I take it as a strategy from 2021–2031, also reflected in the title of this article.
Backdrop for this new national AI strategy
The strategy mentions that it builds on the 2017 Industrial Strategy and the 2018 £1 billion AI Sector Deal that both boosted the UK’s global position in the field of AI. It must be said that there are various other strategies and initiatives that complement this, such as the AI Roadmap by the UK AI Council.
Three assumptions behind strategy
The strategy states in the executive summary that it makes three assumptions. I have shortened these slightly for brevity:
- Key drivers are access to people, data, compute and finance;
- AI will become mainstream in much of the economy effort must be taken to ensure value in different sectors and regions;
- Governance and regulatory regimes of AI will need to keep pace.
Three aims within the strategy
The UK’s National AI Strategy aims to [with added bold and italics]:
- Invest and plan for the long-term needs of the AI ecosystem;
- Support the transition to an AI-enabled economy across all sectors and regions;
- Ensure the UK gets the national and international governance of AI technologies right to encourage innovation, investment, and protect the public and our fundamental values.
Wow this strategy has action points
It was refreshing to see actual action points in a national AI strategy.
Unlike several other national strategies this strategy has action points. This is far more structured than most other national AI strategies that I have read.
You can access the strategy for a detailed overview of the action points that looks like this:
I would like to highlight a few points from the various categories I found interesting.
Short term (next 3 months):
- Develop an all-of-government approach to international AI activity;
- Draft National Strategy for AI-driven technologies in Health and Social Care, through the NHS AI Lab;
- Publish the Defence AI Strategy, through the Ministry of Defence.
Medium term (next 6 months):
- Implement the US UK Declaration on Cooperation in AI R&D;
- Roll out new visa regimes to attract the world’s best AI talent to the UK;
- Extend UK aid to support local innovation in developing countries;
- Pilot an AI Standards Hub to coordinate UK engagement in AI standardisation globally;
- Establish medium and long term horizon scanning functions to increase government’s awareness of AI safety.
Long term (next 12 months):
- Undertake a review of our international and domestic approach to semiconductor supply chains;
- Back diversity in AI by continuing existing interventions across top talent, PhDs, AI and Data Science Conversion Courses and Industrial Funded Masters;
- Work with partners in multilateral and multi-stakeholder fora, and invest in GPAI to shape and support AI governance in line with UK values and priorities;
- Work with The Alan Turing Institute to update guidance on AI ethics and safety in the public sector.
Introduction and more to follow…
The UK believes sees being competitive in AI important to amongst other things:
“…shared global challenges such as net zero, health resilience and environmental sustainability.”
It is important to note that this document sets out strategic intent, but that:
“Detailed and measurable plans for the execution of the first stage of this strategy will be published later this year.”
The UK wants to be the best place to live and work with AI, as well as the best place to study AI.
They also mention democratic values as important to their overall strategy:
“By leading with our democratic values, the UK will work with partners around the world to make sure international agreements embed our ethical values, making clear that progress in AI must be achieved responsibly, according to democratic norms and the rule of law.”
Later in the introduction of the strategy it mentions that it will be linked to ‘interconnected work of government including’:
- The Plan for Growth and recent Innovation Strategy;
- The Integrated Review;
- The National Data Strategy;
- The Plan for Digital Regulation;
- The upcoming National Cyber Strategy;
- The forthcoming Digital Strategy;
- A new Defence AI centre;
- The National Security Technology Innovation exchange (NSTIx);
- The upcoming National Resilience Strategy
Opportunities and challenges
The importance of diversity is again stressed in that it has moral, social and economic reasons for why the UK should care. Ensuring that people from different backgrounds can thrive in the new economy; people from different backgrounds are included; diversity can lead to a wider range of AI services.
Artificial general intelligence is mentioned in passing as a challenge.
DeepMind, GraphCore, DarkTrace and Benevolent AI are mentioned alongside the statement that UK has the 3rd highest number of AI companies in the world.
The strategy mentions what the government has invested in:
“The Office for Artificial Intelligence was created as a new team within government to take responsibility for overarching AI policy across government and to be a focal point for the AI ecosystem through its secretariat of the AI Council.”
Before discussing the three pillars it is important to stress that I have only highlighted part of the information in this article and it would be preferable for you to read the strategy yourself to find other points that I might have missed.
Pillar 1: Investing in the long-term needs of the AI ecosystem
Their approach is to build, use and inspire.
Build: Train and attract the brightest and best people at developing AI.
Use: Empower employers and employees to upskill and understand the opportunities for using AI in a business setting.
Inspire: Support all to be excited by the possibilities of AI.
The strategy states that retaining AI talent is difficult. ‘Understanding the UK AI Labour Market’ research was done by the Office for Artificial Intelligence. Generally there was a lack of suitable talent and lack of diversity in the field.
One aspect of this is changes in visa routes:
- A new High Potential Individual route;
- A new scale-up route;
- The new Global Business Mobility visa.
UKRI AI Innovation Programme
It is worth mentioning that UK Research and Innovation (UKRI) has plans related to AI:
“UKRI will support the transformation of the UK’s capability in AI by launching a National AI Research and Innovation (R&I) Programme. […] an inclusive, interconnected, collaborative, and interdisciplinary research and innovation ecosystem”
It has five aims:
- Discovering and developing transformative new AI technologies;
- Maximising the creativity and adventure of researchers and innovators;
- Building new research and innovation capacity to deliver the ideas, technologies, and workforce of the future;
- Connecting across the UK AI Research and Innovation ecosystem;
- Supporting the UK’s AI Sector and the adoption of AI.
International collaboration on Research & Innovation
It is worth mentioning the UK will participate in Horizon Europe. Focus is also given to implementing the US UK Declaration on Cooperation in AI Research and Development.
Otherwise as elsewhere stated it seems there is an intention to include AI more in international trade deals, especially relating to data flows (action point 12 under pillar 1). In pillar 2 they also mention the Global Partnership on AI (GPAI).
Action points for Pillar 1
Within the first pillar there is a summary of 12 action points. They can be lengthy so I made an attempt here at shortening them:
- Launch a new National AI Research and Innovation Programme;
- Lead the global conversation on AI R&D (Horizon Europe, Overseas Development Assistance and US UK Declaration on Cooperation in AI R&D);
- Develop a diverse and talented workforce (upskill, existing interventions, supporting National Centre for Computing Education (NCCE), promoting visa routes);
- Publish a government policy framework for data availability in the wider economy.
- Start consultation on future national ‘cyber-physical infrastructure’ framework.
- Publish a report on the UK’s compute capacity needs to support AI
- “Continue to publish open and machine-readable data on which AI models for both public and commercial benefit can depend.”
- “Consider what valuable datasets the government should purposefully incentivise or curate that will accelerate the development of valuable AI applications.”
- Wider review of our international and domestic approach to the semiconductor sector.
- Evaluate the state of funding specifically for innovative firms developing AI technologies in the UK (Autumn 2022).
- Protect national security through the National Security & Investment Act while keeping the UK open for business with the rest of the world (protecting against hostile foreign investment.
- Include provisions on emerging digital technologies, including AI, in future trade deals.
Otherwise access to data and increasing compute capacity is mentioned (as in many other national strategies). Still there is an explicit mention of a promised policy framework to come:
“4. Publish a policy framework setting the government’s role in enabling better data availability in the wider economy. The government is already consulting on the opportunity for data intermediaries to support responsible data sharing and data stewardship in the economy and the interplay of AI technologies with the UK’s data rights regime.”
As such this list alongside details in the strategy alludes to various threads of activity in the UK AI landscape.
Pillar 2: Ensuring AI benefits all sectors and regions
The wording used here and throughout the report is the “…transition to an AI-enabled economy…” There is a wish to diffuse AI ‘across the whole economy’. In short:
- Supporting AI businesses;
- Understanding AI adoption in organisations;
- AI for government innovation;
- Leveraging public sector capacity.
There is a wish for more commercialised products and services:
“The UK has historically been excellent at developing new technologies but less so at commercialising them into products and services.” (p. 40)
The UK wants to use AI to support mission-led policymaking, health and achieving net zero are mentioned.
- “…the Office for AI will publish research later this year into the drivers of AI adoption and diffusion.”
Office for AI and UKRI will launch a joint programme that will:
- Support identification and creation of opportunities for businesses;
- Create a pathway for AI developers to start companies;
- Ensure responsible AI products together with AI developers;
- Incentivise investors to learn more about new market opportunities in AI.
Intellectual property and public benefit
The IPO published its AI and IP call for views in 2020 and responded March 2021. This work will be further enhanced going forward:
“…copyright areas of computer generated works and text and data mining, and on patents for AI devised inventions, will be launched before the end of the year.”
The UK wants to protect IP and use AI for public benefit.
“Over the next six to twelve months, the Office for AI will work closely with the Office for Science and Technology Strategy and government departments to understand the government’s strategic goals and where AI can provide a catalytic contribution.”
UK has an Innovation Strategy published the 22nd of July (this one has four pillars).
They mention Imagenet, CIFAR-10, MNIST, GLUE, SquAD and Kaggle. Then they argue dataset challenges can be important for innovation missions in AI.
“The government believes that challenges could be created that simultaneously incentivise significant progress in Innovation Missions while rapidly progressing the development in the technology along desirable lines.”
“To this end, the government will develop a repository of short, medium and long term AI challenges to motivate industry and society to identify and implement real-world solutions to the strategic priorities.”
Net zero and climate change
They mention shared international challenges like health, climate change and poverty. Net zero is again mentioned with a reference to the The Prime Minister’s Ten Point Plan for a Green Industrial Revolution. They write:
“There are a range of climate change mitigation and adaptation challenges that fit this description. These include:
-using machine vision to monitor the environment;
-using machine learning to forecast electricity generation and demand and control its distribution around the network;
-using data analysis to find inefficiencies in emission-heavy industries; and
-using AI to model complex systems, like Earth’s own climate, so we can better prepare for future changes.”
Health and AI
Straight away the £250 million investment43 to create the NHS AI Lab in NHSX is mentioned.
The NHS AI Lab is creating a National Strategy for AI in Health and Social Care in line with the National AI Strategy expected to launch early 2022 setting a direction towards 2030.
The National Strategy for AI in Health and Social Care
Dr Hatim Abdulhussein - General Practice Registrar and Clinical Lead to Digital, AI and Robotics in Education…
Public sector as a buyer (purchasing and procurement of AI)
This relates to how AI procurement can benefit citizens, but also about modernising Armed Forces:
“This will include the establishment of the new Defence AI Centre which will champion AI development and use, and enable rapid development of AI projects.”
The Crown Commercial Service (CCS) worked closely with colleagues in the Office for AI and across government during drafting of guidelines for AI procurement.
“This was used to design their AI Dynamic Purchasing System (DPS) agreement to align with these guidelines, and included a baselines ethics assessment so that suppliers commit only to bidding where they are capable and willing to deliver both the ethical and technical dimensions of a tender.”
The CCS is also piloting training workshops on purchasing AI.
Actions for pillar 2
Within the first pillar there is a summary of seven action points that I have shortened slightly:
- Launch a programme as part of UKRI’s National AI R&I Programme to exploit commercialisation interventions (market opportunities for products and services);
- Launch a draft National Strategy for AI in Health and Social Care (expected launch early 2022 with timeframe towards 2030);
- “Ensure that AI policy supports the government’s ambition to secure strategic advantage through science and technology.”;
- Innovation Missions incorporating AI tackling big, real-world problems such as net zero. Complemented by bilateral and multilateral agreements in energy and with ecosystems in developing AI nations;
- Build an open repository of AI challenges with real-world applications;
- “Publish research into the determinants impacting the diffusion of AI across the economy”;
- Publish the Ministry of Defence AI Strategy.
Pillar 3: Governing AI effectively
“[The] Government’s aim is to build the most trusted and pro-innovation system for AI governance in the world.”
They have outlined several points to achieve this:
- Establishing an AI governance framework
- Enabling AI products and services to be trustworthy (assurance tools, meaningful information to regulators).
- Growing the UK’s contribution to the development of global AI technical standards (regulatory compliance);
- Building UK regulators’ capacities to use and assess AI;
- “Setting an example in the safe and ethical deployment of AI, with the government leading from the front”;
- Working with our partners around the world to promote international agreements and standards that deliver for our prosperity and security, and promote innovation that harnesses the benefits of AI as we embed our values such as fairness, openness, liberty, security, democracy, rule of law and respect for human rights.”
Regarding the last point it is stated clearly that:
“The UK will act to protect against efforts to adopt and apply these technologies in the service of authoritarianism and repression.”
They also state they will defend human rights.
What does pro-innovation governance of AI mean in the UK?
In their strategy they describe what ‘pro-innovation governance of AI’ means in three points:
- That the UK has a clear, proportionate and effective framework for regulating AI that supports innovation while addressing actual risks and harms;
- UK regulators have the flexibility and capabilities to respond effectively to the challenges of AI;
- Organisations can confidently innovate and adopt AI technologies with the right tools and infrastructure to address AI risks and harms (public sector leading the way).
Although admittedly these points and their meaning are not very clear. The first point seem to point to the proposed EU AI Act as a framework for high-risk AI. The second point speaks to increasing UK capabilities, and this could mean hiring more people to assist in responding or technical systems for doing so. The third is a clear expression that the intention is for UK government to use innovations in the field of AI.
They make it clear that AI is currently unregulated.
“The UK already regulates many aspects of the development and use of AI through ‘cross-sector’ legislation and different regulators. For example, there is coverage in areas like data protection (Information Commissioner’s Office), competition (Competition & Markets Authority), human rights and equality (Equality & Human Rights Commission). As well as through ‘sector specific’ legislation and regulators, for example financial services (Financial Conduct Authority) and medical products (Medicines and Healthcare products Regulatory Agency).”
A sector-led approach regulating AI?
This strategy outlines several points on why the UK is taking a sector-led approach.
- The boundaries of AI risks and harms are grey, it is constantly changing.
- Use cases for AI, and their wider impacts, can be highly complex in their own right. They argue for limitations on cross-cutting legislation.
- Individual regulators and industries are already starting to respond to the risks of AI. They argue in the strategy that: “…enabling and empowering individual bodies to respond is a much quicker response to individual harms than agreeing to an AI regulatory regime that makes sense across all sectors.”
- AI is not the only ongoing technology change, they argue for a focus on use cases.
They also however admit that challenges have surfaced and I have shortened these into more condensed bullet points:
- Inconsistent or contradictory approaches across sectors;
- Overlap between regulatory mandates, creating uncertainty about responsibility;
- AI regulation could become framed narrowly around prominent, existing cross-cutting frameworks (data protection), while risks are broader;
- Multilateral, multi stakeholder fora internationally, and global standards development organisations could overtake a national effort to build a consistent approach.
They ask the question in the strategy whether current efforts are adequate. They look at a few alternative options:
- “Removing some existing regulatory burdens where there is evidence they are creating unnecessary barriers to innovation.
- Retaining the existing sector-based approach, ensuring that individual regulators are empowered to work flexibly within their own remits to ensure AI delivers the right outcomes.
- Introducing additional cross-sector principles or rules, specific to AI, to supplement the role of individual regulators to enable more consistency across existing regimes.”
The strategy is not certain regarding these points, yet argues that they must be explored.
Capacity to assess AI
Together with the The Alan Turing Institute the government is exploring what AI capacities that exist in monitoring and assessing products and services. Especially for cross-sectoral AI systems.
They mention Digital Regulation Cooperation Forum, a recently formed voluntary forum comprising the Competition & Markets Authority (CMA), Financial Conduct Authority (FCA), Information Commissioner’s Office (ICO) and Office of Communications (Ofcom) to deliver a joined up approach to digital regulation.
The UK will support international efforts in developing technical standards.
They will pilot an AI Standards Hub to expand the UK’s international engagement and thought leadership.
Additionally they will develop an AI standards engagement toolkit to guide multidisciplinary UK stakeholders to engage in the global AI standardisation landscape.
They also engage with:
- Open Community for Ethics in Autonomous and Intelligent Systems (OCEANIS);
- OECD’s Network of Experts Group on Implementing Trustworthy AI.
UK also contributes to ISO/IEC.
The UK is exploring how to assure compliance with standards and regulations.
To support the development of a mature AI assurance ecosystem, the CDEI is publishing an AI assurance roadmap. In addition to this:
“…the Central Digital and Data Office (CDDO) is conducting research with a view to developing a cross-government standard for algorithmic transparency in line with the commitment in the National Data Strategy.”
Actions for Pillar 3
Again I will try to give you the actions points in a somewhat shortened format:
- Develop a pro-innovation national position (White Paper, early 2022);
- Publish the CDEI assurance roadmap;
- Pilot an AI Standards Hub;
- Continue our engagement to help shape international frameworks;
- Support the continuing development of new capabilities;
- “Publish details of the approaches which the Ministry of Defence will use when adopting and using AI”;
- “Develop a cross-government standard for algorithmic transparency”;
- Work with The Alan Turing Institute to update the guidance on AI ethics and safety in the public sector.
- Coordinate cross-government processes to accurately assess long term AI safety and risks (technical expertise and infrastructure).
- “Work with national security, defence, and leading researchers to understand how to anticipate and prevent catastrophic risks.”
Quantitative indicators for this strategy
Apparently, as a cliffhanger of sorts, the strategy then alludes to indicators that it will use to track this strategy that will be published separately:
“We will publish a set of quantitative indicators, given the far-ranging and hard-to-define impacts AI will have on the economy and society. We will publish these indicators separately to this document and at regular intervals to provide transparency on our progress and to hold ourselves to account.”
According to this last section the recently established Office for Science and Technology Strategy, National Science and Technology Council and National Technology Adviser will also contribute to strategic efforts.
So… That was the 62-page UK National AI Strategy.
Immediate Thoughts on the UK National AI Strategy
I have three immediate thoughts on the strategy without to intricate reflection
Great to see climate change and net zero ambitions included while sustainability could be more explicitly explored
Again it is a bit disappointing to see another AI strategy that does little to discuss sustainability. The word ‘sustainability’ is mentioned once in the introduction never to be repeated. ‘Sustainable’ is mentioned twice. Then again several topics related to the sustainable development goals are discussed.
The strategy is released so close to the COP26 conference, so one would hope that sustainability featured more prominently in the UK AI strategy.
Why was this almost ignored and the timing not considered?
More strategic work on this is needed to ensure sustainable artificial intelligence.
It does not matter much whether we can trust AI systems if we keep reducing the biological diversity on our planet while disregarding growing social inequalities and we may need a more concerted effort on this front.
On the other hand climate change is mentioned and there are related points:
- “The Prime Minister’s Ten Point Plan for a Green Industrial Revolution highlights the development of disruptive technologies such as AI for energy as a key priority
- Government’s Ten Tech Priorities to use digital innovations to reach net zero, the UK has the opportunity to lead the world in climate technologies, supporting us to deliver our ambitious net zero targets.
- This will be key to meet our stated ambition in the Sixth Carbon Budget, and with it a need to consider how to achieve the maximum possible level of emissions reductions”
Optimistic numbers are mentioned too:
“A study by Microsoft and PwC estimated that AI can help deliver a global reduction in emissions of up to 4% by 2030 compared to business as usual, with a concurrent uplift of 4.4% to global GDP. Such estimates are likely to become more accurate over time as the potential of AI becomes more apparent.”
Yet little regard for the life-cycle perspective of digital solutions is included or referred to in this strategy.
It makes me wonder… When will this be addressed properly by the UK?
Where is the money?
As TechCrunch pointed out there is a lack of new money announced to back up the strategy:
“Notably there’s a lack of new money being announced to back up the strategy. […] But there’s no word on how much funding the government might put into supporting the development of AI from here on in…” Natasha Lomas
Many I have discussed the UK AI Sector deal with found it useful that funding was announced in the strategy to back up the action points, yet this is unclear in this new strategy.
Only the US agreement with UK is referred to in this document, and one would hope the UK had more clear plans for other partnerships going forward. Of course the UK collaborates with a wide range of countries to the listed initiatives, yet perhaps other regions could be mentioned more prominently.
Especially in light of the recent disagreements with France regarding the AUKUS agreement one would hope this strategy was more integrated with nations in Europe that have strong AI communities. This is equally important for Germany, Spain etc. It would perhaps make sense for UK to attempt to reintegrate with Europe to a greater extent also in their efforts within the field of AI.
Is this a gift to artificial intelligence and technology policy wonks?
The clear answer is of course yes!
This is one of the most comprehensive national AI strategies that I have ever seen (and I have read most).
Policy wonk: “a person who takes an enthusiastic or excessive interest in minor details of political policy.”
So perhaps the answer was obvious from the start?
Then again we should be asking:
- Is this a gift to the planet?
The answer to this question is not clear, with a stretch the answer could be ‘maybe’.
It is positive that the strategy embraces climate change, net zero and human rights.
It is less fortunate that it does not map out or include to any great extent the sustainable development goals.
What do you think?
Overall I think this is one of the better national strategies I have read in the field of AI, and it remains to be seen how it will be executed.
I hope this was useful and please feel free to leave a comment or message me directly on Twitter.
This is #1000daysofAI and you are reading article 506. I am writing one new article about or related to artificial intelligence for 1000 days. The first 500 days I wrote an article every day, and now from 500 to 1000 I write at a different pace.