AI Barometer 2021
The CDEI has published the second edition of its AI Barometer, alongside the results of a major survey of UK businesses.
Documents
Details
The CDEI’s AI Barometer is a major analysis of the most pressing opportunities, risks and governance challenges associated with AI and data use in the UK. This edition builds on the first AI Barometer and looks at three further sectors - including recruitment and workforce management, education, and transport and logistics - that have been severely impacted by the COVID-19 pandemic. Alongside the AI Barometer, the CDEI has published the findings from a major survey of UK businesses across eight sectors, which it commissioned Ipsos MORI to conduct to find out about businesses’ readiness to adapt to an increasingly data-driven world. The survey was undertaken between March and May 2021, with analysis from Enterprise Research Centre (ERC).
To produce the AI Barometer, the CDEI conducted an extensive review of policy and academic literature, and convened over 80 expert panellists from industry, civil society, academia and government. It used a novel comparative survey to enable panellists to meaningfully assess a large number of technological impacts; the results of which informed a series of workshops.
Key findings
-
The CDEI’s analysis highlights the ‘prize to be won’ in each sector if data and AI are leveraged effectively. It points to the potential for data and AI to be harnessed to increase energy efficiency and drive down carbon emissions; improve fairness in recruitment and management contexts; and enable scalable high quality, personalised education, in turn improving social mobility.
-
In order to realise these opportunities, however, it points to major, cross-sectoral barriers to trustworthy innovation that need to be addressed, ranging from a lack of clarity about governance in specific sectors, to public unease about data and AI are used.
-
The report also ranks a range of risks associated with AI and data use. Top risks common to all sectors examined include: bias in algorithmic decision-making; low accuracy of data-driven tools; the failure of consent mechanisms to give people meaningful control over their data; as well as a lack of transparency around how AI and data is used.
-
The major survey of UK businesses generated a range of insights. For example, it revealed that businesses that have extensively deployed data-driven technologies in their processes would like further legal guidance on data collection, use and sharing (78%), as well as subsidised or free legal support on how to interpret regulation (87%).
Next steps
The CDEI’s work programme is helping to tackle barriers to innovation. It worked with the Recruitment and Employment Confederation to develop industry-led guidance to enable the responsible use of AI in recruitment. It has also set out the steps required to build a world-leading ecosystem of products and services that will drive AI adoption by verifying that data-driven systems are trustworthy, effective and compliant with regulation. Separately, it is working with the Cabinet Office’s Central Digital and Data Office to pilot one of the world’s first national algorithmic transparency standards, as well as with the Office for Artificial Intelligence as it develops the forthcoming White Paper on the governance and regulation of AI.