A framework developed by researchers at the Alan Turing Institute titled ‘the Human Rights, Democracy, and the Rule of Law Risk and Impact Assessment for AI Systems (HUDERIA)’ has been adopted by the Council of Europe.
The HUDERIA methodology is a multi-step governance process that provides an evidence-based and structured approach to carrying out risk and impact assessments for AI systems.
It is designed to help both public and private organisations identify and mitigate adverse impacts to human rights, democracy and the rule of law.
HUDERIA provides an anticipatory approach to the governance of AI design, development, and deployment that helps users protect the public from potential harms caused by artificial intelligence.
For instance, if an AI system used in recruitment is found to pose significant risks of bias against certain demographic groups, the mitigation plan might involve resampling training data or adjusting the algorithm to address discriminatory proxies and then testing the system to ensure sufficient bias mitigation.
The HUDERIA framework contains four elements; a context based risk analysis, a stakeholder engagement process; a risk and impact assessment and a mitigation plan.
Using the framework also prompts regular reassessments to ensure that AI systems continue to operate safely and ethically as the context and technology evolve, ensuring that the public is protected from emerging risks throughout the system’s life cycle.
Researchers from the Turing’s Ethics and Responsible Innovation team have been developing HUDERIA since 2020, working closely with the Council of Europe and its Member and Observer States.
Turing contributions, led by Director of Ethics and Responsible Innovation Research David Leslie, have included:
- Publishing Human Right, Democracy, and the Rule of Law: A Primer in 2021
- Carrying out a study on what an impact assessment instrument to support a future treaty could look like in Human Rights, Democracy and Rule of Law
- Assurance Framework for AI systems: A proposal which formed the basis of HUDERIA.
- Supporting the drafting of the Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law and the HUDERIA methodology.
- Undertaking a first-of-its-kind initiative to pilot HUDERIA with a group of Council of Europe Member and Observer States.
The HUDERIA methodology was adopted by the CAI at its 12th plenary meeting, held in Strasbourg last week. The plenary also set out next steps for the project, agreeing that in 2025-2026 a more detailed model will be developed and piloted.
Professor David Leslie said: “This is a historic achievement for both the Council of Europe and the Alan Turing Institute. It represents a quantum leap in the advancement of rights-based approaches to international AI policy and law at a critical juncture when much of the future of global AI governance regimes remains undecided.
"We are honoured to play a part in supporting this momentous effort and look forward to supporting the next iteration of the HUDERIA, offering a global-facing good practice resource and library of knowledge.”
Top image credit: Kittyfly via Adobe Stock
