18 March 2021

How Britain can lead the world in developing artificial intelligence

By Seb Krier

From diagnosing diseases to driverless cars, artificial intelligence has the potential change the way we live for the better. The UK has been a driving force in its development, and was one of the world’s first countries to announce policies to strengthen the sector through its AI Sectoral Deal back in 2018. However, since then, not much has happened. Theresa May’s short-lived industrial strategy produced questionable results, and the civil service has been distracted by Brexit negotiations and the pandemic.

As a result, with some notable exceptions, there have been very few concrete initiatives and policies on AI – mostly expressions of intent, panel talks and some guidance. The European Union, on the other hand, is set to announce a host of programmes and regulations in April further to their AI White Paper.

Last week, and partly in response to the AI Council’s AI Roadmap, the Government announced their intention to publish a National AI Strategy later this year. This is encouraging but a lot will depend on the detail and the approach the Government takes. In a report for The Entrepreneurs Network, I’ve set out eight ways to make the UK the best place in the world for AI.

Unlocking Talent

First, we need to create a pool of cloud computing resources for the UK AI researchers. This is important because AI R&D is dependent on prohibitively expensive hardware, making it more difficult for academics outside the private sector to undertake more ambitious projects. Such an initiative, if designed carefully and without excessive government involvement, could provide leading academics with the possibility of undertaking ambitious research without having to necessarily move to large tech firms.

We also need to lower barriers to immigration and attract foreign talent. Boris Johnson’s speech in Greenwich set out a vision for an open society and thriving markets, which is laudable but it can only be realised with the right skills. The UK however is facing a skills shortage: only 17% of UK employees are being re-skilled for AI, compared to 38% globally. Because government-funded masters and PhD programmes will take many years to bear fruit, it’s crucial to attract as many international students, entrepreneurs, and high-skilled workers as we can. The recently announced Turing scheme and the revamped visa system are positive steps, but more should be done to liberalise the immigration system as a whole.

Open Data and Better Regulation

I mentioned above the EU’s AI regulation plans; while there are doubts as to the efficacy of some of the proposed measures, this remains important work and the UK should make sure they do not fall behind. The UK could differentiate itself by adopting a more innovative approach to AI oversight: regulatory markets. This idea, authored by Jack Clark and Gillian Hadfield, proposes a model where the government creates a market for regulation in which private sector organisations compete to achieve regulatory outcomes set by a government regulator. This allows flexibility but also ensures there are market incentives to create tools to address risks associated with AI systems.

The UK also needs to work closely with the EU to review and improve GDPR. Axel Voss, one of the fathers of GDPR, warns that “We have to be aware that GDPR is not made for blockchain, facial or voice recognition, text and data mining [ . . . ] artificial intelligence.” Privacy is important and should not be watered down, but the appropriate balance needs to be struck to ensure both the EU and the UK do not fall behind China and the US. For example, this should include working toward reducing fragmentation and allowing repurposing of data that poses only minimal risk.

GDPR is not the only piece of legislation in need of an update: the intellectual property landscape in the UK does not permit many uses of text and data mining. In a recent report, Dr Anton Howes recommends extending the exemption from copyright for text-and-data mining to for-profit uses. With the law as it currently stands, companies are reticent to share details of how they train their algorithms – improving the IP regime might help them become more transparent, for example as part of AI audits.

The UK used to be particularly proactive in upgrading public data infrastructure and opening up datasets, but unfortunately progress has stalled in recent years. The Open Government Network recently urged the UK to improve its open government agenda, further to the country being placed under review by the Open Government Partnership for failing to meet the required standards. Opening up datasets, APIs and particularly real-time data will enable a lot more innovation, but also enhanced transparency.

Global Leadership and Governance

For the UK to continue leading globally, it should resist the temptation to give in to digital protectionism, and support standards that promote openness, interoperability and competition. Policies advocating for preferential treatment, attempts to bypass state aid rules, unequal access to data and state-grown champions should be resisted. In addition, the UK should collaborate with allies to pool technological resources, defend against digital authoritarianism, and advocate for inclusive growth, human rights, and democratic values. For example, the newly announced Advanced Research and Invention Agency could devote part of its work to developing open and accessible technology-focused projects that promote human rights, like the US Open Technology Fund does.

Lastly, the UK government should not ignore or discount concerns around AI safety and ethics. AI presents new challenges and uncertainties, and it is legitimate for the public to have some concerns on how it will be developed and used. In the US, minorities have been wrongfully accused by error-prone facial recognition systems. To ensure AI is both trusted and trustworthy, the Government should reform the public institutional landscape. A lot of the Centre for Data Ethics and Innovation [CDEI] outputs, while of high quality, are already undertaken by other organisations: instead the CDEI seems better positioned to scrutinise the public sector use of AI. It could also help cities trial AI registers. The Office for AI, on the other hand, needs a proactive mandate beyond drafting strategies: it would benefit from more funding, a clearer delineation of scope and responsibilities, and a more policy-oriented agenda like the US OSTP.

AI will have important implications for the environment, productivity and competition. With the bulk of Brexit negotiations behind us and commendable progress on vaccinations, now is the perfect time for the UK to announce bold, impactful new ideas. AI is a relatively new field but innovation should not be limited to labs – with the right policies and incentives Britain can lead the world.

Click here to subscribe to our daily briefing – the best pieces from CapX and across the web.

CapX depends on the generosity of its readers. If you value what we do, please consider making a donation.

Seb Krier is former Head of Regulation at the Office for AI.

Columns are the author's own opinion and do not necessarily reflect the views of CapX.