Effective Remedies in AI: An Insufficiently Explored Avenue for AI Accountability
The right to an effective remedy, or effective redress, enshrined in the European Charter of Fundamental Rights, plays a crucial role in operationalizing all rights in the EU — whether fundamental or not. It ensures that the effective enforcement of legally-backed rights in the EU is a free-standing fundamental right, holding EU institutions and member states accountable to individuals when implementing EU law.
This foundation is vital for ensuring that protections set in legislation governing the digital sphere, including the AI Act, are effectively applied. However, the unique challenges presented by AI and other complex technologies complicate access to effective remedies.
The Obstacles to Effective Remedies in AI
AI presents a novel challenge to effective remedies due to the general lack of information regarding how AI systems function. This opacity and the frequent lack of disclosure regarding AI usage hinder individuals’ awareness of how they may have been affected by an AI system.
Legal procedures often struggle to keep pace with the reality of digital products. Traditional rules concerning the burden of proof require victims to explain the internal workings of complex systems that often operate with significant opacity and autonomy. The European Commission’s Digital Fairness report highlights that these obstacles can render the right to compensation practically ineffective, even when complainants possess sufficient information.
The AI Act addresses the transparency issue to some extent: providers deploying AI systems directly interacting with individuals are required to disclose the use of such systems unless it is obvious, and individuals must be notified if they are subjected to decision-making supported by a high-risk AI system. While these steps are commendable, transparency alone does not guarantee effective remedies.
Recognizing the necessity for a dual approach that addresses both foundational transparency challenges and procedural difficulties, the European Commission began exploring a liability framework for AI-related harms as early as 2020. This exploration coincided with the development of the white paper leading to the AI Act, resulting in the proposal known as the AI Liability Directive (AILD), which aims to ease procedural burdens for complainants faced with the hurdles posed by AI’s opaque functioning. Unfortunately, the AILD has struggled to gain traction, leaving the effective remedies issue in AI largely unaddressed.
The AI Act’s Lukewarm Approach to Individual Rights
The parallel development of the AI Act and the AILD indicates that the AI Act was not intended to tackle the effective remedies issue independently. Instead, it aims to bolster the protection of fundamental rights in the AI era.
Within its sparse “Remedies” section, the AI Act proposes two rights: the right to obtain an explanation for any decision made based on the output of a high-risk AI system with significant effects, and the right to lodge a complaint with a market surveillance authority regarding any infringement of the AI Act. Notably, anyone, whether an individual or an entity, can file a complaint, expanding the scope of potential complainants. However, this expansion may come at the expense of thorough assessment and processing of complaints.
Critically, the AI Act does not compel market surveillance authorities to engage substantively with complaints or the complainants themselves. This lack of obligation renders the complaint mechanism established by the AI Act weak, contrasting sharply with the higher standards set by data protection legislation such as the General Data Protection Regulation and the Law Enforcement Directive, which guarantee individuals an effective judicial remedy against supervisory authorities’ binding decisions.
Consequently, individuals are left with no assurance of meaningful engagement with the regulators under the AI Act. This inadequacy raises concerns regarding the accountability of national AI regulators and, by extension, AI providers and deployers, if they fail to act appropriately. Overall, the right to complain to national AI regulators lacks sufficient enforcement.
The Need for an Ongoing Focus on the Effective Remedies Issue
The discussion surrounding effective remedies in AI is both necessary and urgent, especially as pro-innovation narratives gain traction within the European Union. Current trends suggest a pro-industry stance that conflicts with the expressed desire to uphold fundamental rights. As emphasized by the European Charter of Fundamental Rights and the jurisprudence of the European Court of Justice, the right to effective remedies is not optional.
Future efforts must focus on exploring the various ways existing and proposed laws in the EU, beyond the AI Act, can enhance access to effective remedies for AI-induced harms, identifying and addressing any existing gaps.
