On the proposal for a directive of the European Parliament and of the Council on adapting non-contractual civil liability rules to artificial intelligence (AI Liability Directive)

The adoption of an AI Liability Directive at this stage is premature and unnecessary.

The AI Act, which entered into force on 1 August 2024, introduces comprehensive regulations on data transparency, risk management, and due diligence. Depending on the risk classification, these rules will be applied by August 2026 or 2027, significantly transforming the AI regulatory framework in Europe. Additionally, the newly adopted Product Liability Directive (EU) 2024/2853, which must be transposed by the end of 2026, has already expanded liability rules for software and AI applications. That Directive imposes stricter compliance requirements, making it easier for individuals to claim damages for harm caused by AI. Moreover, national tort laws across Member States provide additional avenues for liability claims. The combined effect of these rules represents a fundamental shift in the legal systems governing AI, and it will take several years before we can fully understand their real impact on the internal market, including innovation, Europe’s global competitiveness in the AI sector, manufacturing, consumer welfare, businesses, economic growth, or societal development.

In its State of the Digital Decade 2024 report, tracking the yearly progress towards the targets set for 2030 in the Digital Decade Policy Programme, the European Commission reported that “The percentage of enterprises using AI increased by a mere 0.4 percentage point in 2 years, from 7.6% in 2021 to 8.0% in 2023” . The same report concludes that, at this pace, only 16.8% of enterprises will take up AI by 2030 compared to an EU target of 75%. Against this background, it is essential to take innovation and uptake-friendly measures rather than increasing the liability burden on businesses.

In his 2024 report, Mario Draghi stated “we claim to favour innovation, but we continue to add regulatory burdens onto European companies, which are especially costly for SMEs and self-defeating for those in the digital sectors”. He also provided a stark warning that the bloc is falling massively behind and losing the economic fight with the global players — and that huge investment and regulatory overhaul is urgently needed. Adding another layer of regulation would undermine the EU’s goal of fostering AI innovation and attracting AI businesses and talent back to Europe, as emphasized also by Commission President Ursula von der Leyen, the newly appointed commissioners and various strategy papers.

From the perspective of better regulation, introducing an AI Liability Directive (AILD) now contradicts the principles of evidence-based policymaking. The European Commission’s 2022 impact assessment for the proposal (‘the AILD impact assessment’) relied on hypothetical scenarios rather than concrete data without taking into account the final texts of the AI Act and the revised Product Liability Directive. Indicative of this is the tone in which the AILD impact assessment disregards costs increase in insurance premiums as ‘not a significant burden’ while also admitting the technologies the AILD seeks to regulate are not widely on the market and no statistical data on the damage caused by these products and services exists.

Under Art. 96(1) of the AI Act, the European Commission is due to publish guidelines on the definition of “Artificial Intelligence Systems”, which the proposed AILD is also based on. Such guidelines do not exist yet and are crucial for businesses to fully understand when their technology falls under the definition of AI system. It would be against better regulation principles to work on an AILD proposal when its very subject is still not fully clear. Moreover, a number of items of secondary legislation are expected to clarify many parts of the AI Act obligations and even the scope of the high-risk classification (eg. guidelines on the relation with Union harmonisation legislation in Annex I), which also plays a role in the proposed AILD. These uncertainties undermine the validity of the cost-benefit analysis underpinning the AILD proposal, as such analyses lack the necessary information to be accurate or reliable.

An excessive liability or rather procedural law framework risks deterring innovation and increasing compliance burdens, particularly for SMEs. This proposal focuses less on liability and more on procedural rules, namely evidence disclosure and burden of proof. Such changes could disrupt national civil law systems, which have functioned effectively for decades, including in the digital era. The Commission and scholars fail to provide any market data or analysis on the real-world impact of the proposal. As a result, related papers present more of a theoretical, academic perspective rather than a balanced assessment grounded in reality and backed by a comprehensive consideration of relevant data.

The harmonisation argument appears speculative, and further clarification is needed to substantiate this claim. Without clear evidence of market failures or difficulties applying existing liability rules to digital technologies, new legislation risks creating unnecessary complexity.

Given the extensive new obligations introduced by the AI Act and the revised Product Liability Directive, it is prudent to wait for their implementation and evaluate their real internal market impact. For the effects of the AI Act itself to be known, time will be needed. As pointed out by the European Court of Auditors, the AI Act impact assessment failed to provide evidence on how attractive the rules would be for enticing investment into the EU, and the regulatory costs borne by investors and the EU’s competitive position will depend on implementation.

At this stage, there is no demonstrated legal gap or market failure justifying a separate AI Liability Directive. The existing frameworks might be sufficient to address AI-related liability issues. Any theoretical benefits of this proposal are outweighed by the fact that the AILD does not reallocate liability but simply makes (predatory) litigation, especially third-party litigation funding, easier. To ensure a stable investment climate, third party funding must be addressed before any more progress on AILD is made. We should avoid overregulation and allow the new legislative packages to be implemented and tested in practice. Only after collecting real market data and conducting a thorough impact assessment including competitiveness check, SMEs check and an evaluation of its impact when accumulated with other recent digital legislation should additional, targeted measures be considered. This approach aligns with the EU’s commitment to better regulation and innovation-friendly policies.

******

The Committee on the Internal Market and Consumer Protection calls on the Committee on Legal Affairs, as the committee responsible, to propose rejection of the Commission proposal.

Pursuant to Article 8 of Annex I to the Rules of Procedure, the rapporteur for the opinion declares that he received input from the following entities or persons in the preparation of the draft opinion:

The list above is drawn up under the exclusive responsibility of the rapporteur for the opinion.

Where natural persons are identified in the list by their name, by their function or by both, the rapporteur for the opinion declares that he has submitted to the concerned natural persons the European Parliament's Data Protection Notice No 484 (https://www.europarl.europa.eu/data-protect/index.do), which sets out the conditions applicable to the processing of their personal data and the rights linked to that processing.