Algorithmic Impact Assessments

Algorithmic systems are increasingly used by Governments and businesses to make ‘automated’, data-driven decisions that are having far-reaching consequences on work and working people. Existing regimes for regulation have been outpaced, requiring new proposals to ensure meaningful accountability and shape innovation in the public interest.

This policy paper and toolkit proposes algorithmic impact assessments as a systematic framework of accountability for algorithmic decision making to support the next phase of the AI Strategy.

Impact

• We have worked closely with the Canadian Government on algorithmic decision-making which has informed our publication of our prototype for new, affirmative ex ante legal duties. This work was presented in a workshop convened for the Office for AI and fed into their consultation process ahead of the publication of the UK’s AI White Paper.
• Informed an All-Party Parliamentary Group on the Future of Work APPG event on ‘International Learnings on Algorithmic Impact Assessments’, featuring contributions from the Canadian Government, and written up in Lexology and on the Burges Salmon blog.
• The report has been featured in a list of frameworks, guidelines and toolkits on AIethicists.org.