Image: © peepo | iStock
AI is a Westminster buzzword to tackle many of the UK’s systemic challenges, but the government must hone its use among local authorities first, write Anja Beriro and James Arrowsmith, partners at UK and Ireland law firm Browne Jacobson
Barely a UK Government report is published without some mention of artificial intelligence (AI) for driving efficiencies or innovation.
Papers like the ‘AI Playbook for the UK Government’ and the ‘AI Opportunities Action Plan’ also aim to set the path for how Britain can become a global leader in responsibly embracing AI.
Such lofty national ambitions, however, appear to neglect the fact that local government holds the keys to many wings of the AI castle.
Devolution and AI
Westminster wants to create ‘AI Growth Zones’ to accelerate the construction of AI data centres, as outlined in the ‘AI Opportunities Action Plan’. Significant hurdles to overcome relate to planning, funding, and infrastructure limitations, such as connecting sites with sufficient clean energy.
Widening and deepening devolution across England means it’s not just central government but also mayoral combined authorities and councils that can support the delivery of data centres.
Devolved skills budgets also give them responsibility for upskilling workers in AI and associated roles, as well as being key stakeholders in the UK’s 12 freeports. These aim to create hubs for innovation and trade across industries such as clean energy, life sciences, logistics and manufacturing via tax and customs incentives.
Tapping into local knowledge is imperative. Whitehall may have concerns over the potential for local opposition to data centre development, but planning officers have a key role.
Developments that are locally led, aligned to placemaking strategies, and integrated with plans to boost local economies, jobs and training are more likely to achieve local buy-in.
Centralising data sets
With a public sector that employs about 18% of the UK’s total workforce, and directly engages in some way with every citizen, there is huge untapped potential to leverage the data held across local and central government, education and the NHS.
An ambition in the ‘AI Opportunities Action Plan’ to create a ‘National Data Library’, therefore, brings significant potential to unlock data assets in both the public and private sectors.
But while public services are deeply integrated into people’s lives, Britain has historically been very poor at collecting and analysing data in a consistent and coherent manner.
Government started this journey by publishing guidance in January on how public bodies can make government datasets ready for AI, including a self-assessment checklist.
However, it must prioritise datasets with the highest potential to enable AI to drive significant positive change, identify who owns that data, and lead rapid work on standardisation and sharing.
The commercial worth of existing and future datasets must be recognised from the outset, to secure these assets and find appropriate opportunities for public sector revenue generation.
Walking the tightrope between responsible AI policy and innovation
With the ‘EU AI Act’ arguably taking a safety-first approach to AI and the U.S. dogmatically pro-innovation, there’s an opportunity for the UK to take a leadership role in responsible AI policy adoption that avoids stifling innovation if it can successfully walk this regulatory tightrope.
To achieve this, all parts of the public sector must develop the capability for sophisticated risk management, balancing opportunity and hazard risks within robust frameworks that reflect a considered risk appetite.
For safe innovation to happen, a significant improvement is required in how we understand risk, innovation and opportunity.
The government’s role is to take a clear and honest appraisal of the current system and its weaknesses. It must then clearly express priorities and risk appetite in key areas of AI innovation.
For example, using AI for identifying potholes and school lesson planning will clearly have a lower hazard risk than detecting cancer, but the greater opportunity is likely to lie in disease diagnosis and prevention. AI choices may sometimes need to be made not because they are easy, but because they are hard.
Training public sector workers will be crucial to safe AI adoption, as a recent mishap involving West Midlands Police and AI-generated false intelligence illustrated. This not only led to a poor decision, but also contributes to undermining trust in institutions and the technology they use – trust that is needed as AI use expands.
This demonstrates why public bodies must establish AI frameworks that cover risk management, data governance, information security protocols, and internal expertise for effective oversight of AI vendors.
AI’s transformative potential
Despite the risks involved with AI, successful adoption brings remarkable potential to transform how public services interact with their communities.
An exciting example can be found in Waves, a collaboration between Google, think tanks Demos and New Local, Camden Council, and South Staffordshire District Council. The project is testing how AI can make it easier, cheaper and quicker for residents to have a say in tackling contentious local issues.
By speedily identifying areas of consensus, where difficult issues remain and how trade-offs can be mutually agreed, the idea is to improve engagement and trust in local institutions via mass deliberative democracy. It also supports public bodies in gaining greater insight into citizens’ needs and priorities to inform decision-making.
When we consider how this might be applied to difficult decisions around care services or infrastructure development, there is hope that using AI to bring people along on the journey could have a positive impact on delivering a well-functioning society and economy.
AI could even be a tool for building consensus around the delivery of its own data centres.