The AI Council has urged the Government to give civil servants a central source of information on AI based tools and projects and to fund pathfinder projects to boost the public sector take-up of the technology.
It has sounded the call in its new AI Roadmap report, along with recommendations to promote usage in health and social care and for a range of measures relevant to the broad economy.
The council, an independent expert committee set up by the Government, says the public sector should be a priority area for AI innovation, which could be harnessed to improve services and policy making.
An important element of this would be to ensure relevant data can be accessed and evaluated by public servants, which prompts the proposal for a central place to find information on tools and projects.
The report says AI should only be used in projects that are outcome based and that departments should learn rapidly from each other about what works and what needs to evolve. Over time, there should be a focus on linking public and private sector data to improve insights and encourage a sophisticated adoption of AI.
This should be accompanied by the Government funding pathfinder projects that involve public authorities and industry working together, partly with the aim of building the capability to deliver within the Civil Service.
There is also a need to demystify algorithmic decision making, build up data science skills in government, and to support intelligent procurement of AI products, the report says. This would involve working with the Alan Turing Institute, the Royal Society and the Royal Statistical Society.
This comes with a reiteration of the need to ensure the technology is applied ethically, and a recommendation that independent committees are given the job of considering the ethical, social and economic impacts of its use.
For health and social care, the roadmap says AI could be used to manage complex information flows and the integration of systems, between different elements of care. This would have been valuable in the response to the Covid-19 pandemic if the sector had taken up AI earlier.
The council says that NHSX should work with other bodies on research and innovation, and that extending the use of AI beyond hospitals will increase the prospects for effective preventative care, and give the UK economy a competitive advantage in the field.
Underlying this is a need for a data strategy for health and care, new models for incentives and partnerships between the NHS and SMEs, and a better understanding of the skills and knowledge required by healthcare professionals and the public.
On a broad front, the report looks at the development of AI in the UK with its implications for the economy, saying it could deliver a 10% increase in GDP in 2030, and emphasises two underlying messages.
“The first is that we need to ‘double down’ on recent investment the UK has made in AI,” it says.
“The second message is that we must look to the horizon and be adaptable to disruption. Support for AI needs to reflect the rapid pace and evolution of the science and technology and its applications. This means staying at the forefront of the development of AI and integrating approaches to ethics, security and social impacts and planning for the next 10-50 years.”
Its range of general recommendations includes: scaling up public sector investment in the technology; cementing the role of the Alan Turing Institute; the creation of a 10-year programme of building skills at a high level; committing to achieving AI and data literary for everyone; providing for scrutiny to build public trust in how the data is used; developing data governance options; and supporting start-up vendors of the technology.
Image from iStock, Andrey Suslov