As part of Equinet’s innovative work on promoting an equality-compliant use of Artificial Intelligence (AI) in Europe, we present our recommendations for the decisive negotiations which will shape the future of the Artificial Intelligence Act, known as the AI Act (AIA).
AI Act: towards a more equal Europe
The AI Act could be a new frontier in the protection of equality in the age of expanding and diversifying AI uses. As independent watchdogs for equality and a source of empowerment of rightsholders, Equality Bodies have a key role to play in the future implementation of the AIA to ensure better access to justice and protection from discrimination for all.
Equinet has contributed to the work on the AIA from the very start of the process, first through its Report “Regulating for Equal AI: A New Role for Equality Bodies” and subsequently, through consultations in relation to the Commission’s White Paper on AI and upon the release of the European Commission’s proposal. In addition, in July 2022, Equinet issued a Position Paper “Regulating for European AI that Protects and Advances Equality”, which includes recommendations on how to improve the protection of equality in the future horizontal Artificial Intelligence Act (AIA).
Ahead of the start of the Trilogues, the below recommendations seek to strengthen compliance with equality and non-discrimination law, remove barriers to its enforcement and complement it with additional AI-specific safeguards. While we welcome the strengthening of fundamental rights protection by the European Parliament and the inclusion of a complaints mechanism by the European Council, the AI Act needs more robust safeguards for the rights of those affected by AI systems and lacks clarity and legal certainty in relation to the enforcement of equality and non-discrimination obligations.
You can access Equinet’s recommendations in French on the Défenseur des droits’ website.
Equinet’s recommendations
The AIA must ensure access to justice for all rights-holders and strengthen the role of Equality Bodies as a crucial redress mechanism for those affected by AI-enabled discrimination. To that end, the AIA should clearly ensure access to information for Equality Bodies that is practically useable in their work, as well as the ability to rely on the technical assistance of supervisory authorities whenever information is not sufficiently clear.
The AIA must ensure harmonised and consistent enforcement of non-discrimination obligations. It is important to create the conditions for Equality Bodies to share their expertise on non-discrimination and cooperate with supervisory authorities in evaluating and enforcing against discrimination risks, acting on discrimination complaints, and carrying out the impact assessments of AI systems. This would prevent two parallel enforcement mechanisms with divergent or even contradictory findings.
The AIA must ensure that rights holders affected by AI systems can access non-discrimination protection, which grants a series of benefits: the reversal of the burden of proof, the ability of Equality Bodies to act in defense of rights-holders and in their own name, including without an identifiable victim, the right to lodge a complaint and the right to effective judicial remedies.