Source: European Parliament
1. The proposed Artificial Intelligence (AI) Liability Directive[1] was designed with a forward-looking approach to modernise certain fault-based liability rules, considering AI’s unique characteristics (opacity, complexity and autonomy), which differentiate it from traditional software.
Legal certainty regarding liability is important for reducing risks and encouraging businesses, especially small and medium-sized enterprises and startups, to invest in AI. Harmonised rules for fault-based liability minimise fragmentation, creating a predictable environment for innovation while ensuring accountability.
2. The proposed Directive does not create any administrative burden for businesses (no reporting, no registration, no documentation obligation) and enhances legal certainty.
The proposed AI Liability Directive is targeted and proportionate, as it only deals with aspects of liability that are challenged by the specificities of the AI, while all other aspects of liability are left to national law.
3. The Commission’s aim is to create a framework that does not stifle growth. The proposed Directive does not create any additional hurdles for companies, as it relies on obligations that companies must already comply with under European or national law.
The AI Act[2] reduces risks for safety and fundamental rights, the Product Liability Directive[3] sets no-fault liability of producers for defective products, benefiting consumers, and the proposed AI Liability Directive covers liability for damages caused by AI, linked to the fault of any persons.
Adaptation of fault-based liability to the specificities of AI allows any type of victims of accidents caused by AI systems to prove a successful liability claim and obtain compensation.