Public use of AI decision-making needs legal change

Justice Lab has today published a response to the Law Commission’s consultation on its 14th programme of law reform. The response offers strong support for the Law Commission’s proposal to develop a new legal framework to support the use of automated decision-making by public sector bodies.

You can read the full response here.

Since 2018, Justice Lab has developed a focus on the public and human rights law implications of the increasing use of automated and assisted decision-making technologies (“ADM/ASDM”) by public bodies. We have funded organisations working in the access to justice space to develop work to understand and respond to the challenges presented by the rapid expansion in the use of these tools in areas such as welfare benefits and immigration. We have also funded experts in equality law to examine the issues raised by particular uses of automated and assisted decision making in the public sector. In developing our response, we convened a technical legal workshop and commissioned new research from experts Professor Lilian Edwards, Professor Rebecca Williams and Reuben Binns.

Automated Decision Making (“ADM”) and Assisted Decision Making (“ASDM”) systems are widely used to support decision making across the public sector. When deployed appropriately and lawfully, the adoption of these technologies offers the potential to improve the speed and consistency of decision-making whilst generating significant savings for the taxpayer. However, recent experiences in immigration, policing, welfare and education have highlighted the risks and limitations associated with the use of ADM/ADSM systems. These risks, if unaddressed, can undermine rights, damage trust in public sector bodies and generate costly litigation.

Dr Natalie Byrom, Director of Justice Lab

Justice Lab’s response argues that existing legal frameworks are complex, piecemeal and provide insufficient clarity about whether, when or how ADM/ASDM systems can lawfully be deployed by public bodies.

Existing law creates uncertainty and increases the risk that ADM/ASDM systems will be deployed inappropriately, with deleterious consequences for individuals, public bodies and the taxpayer. Mechanisms for seeking redress are both complex and expensive to access, and the actual redress available limited.

Dr Natalie Byrom, Director of Justice Lab

Justice Lab’s response argues that to develop effective solutions, The Law Commission should examine and learn from national international approaches to regulating public sector use of ADM/ASDM systems. In identifying models of best practice on which to base the Law Commission’s approach, primacy should be given to those frameworks that can be demonstrated to:

• Address both ADM and ASDM systems
• Ensure that ADM/ASDM systems uphold existing equalities and human rights law
• Secure meaningful and effective transparency in relation to the use of ADM/ASDM systems e.g. via public registers
• Deliver certainty for public bodies, suppliers and individuals around the circumstances in which ADM and ASDM systems can be used;
• Support meaningful public engagement in determining appropriate uses of ADM/ASDM
• Focus governance at the design and deployment stage
• Are capable of managing and responding to contextual complexity
• Introduce independent external scrutiny to ensure the efficacy and accuracy of ADM/ASDM systems
• Ensure clear lines of accountability for decisions taken by ADM/ASDM systems, and secure legally enforceable rights to an explanation.
• Provide timely, appropriate, accessible and cost-effective routes to redress where this is required.

The Law Commission’s consultation closed on 31st July. You can read more about the consultation here

Further reading

Justice Lab’s response to the Law Commission’s consultation can be found here.

The workshop report: “Reforming the law around the use of automated and assisted decision making by public bodies” can be found here.

Research conducted by Professor Edwards, Professor Williams and Mr Binns: “Legal and regulatory frameworks governing the use of automated and assisted decision making by public bodies” can be found here.