Product liability and AI liability directives - a BusinessEurope position paper

Key messages

• The expansion of scope of the PLD should be limited to embedded software. For instances of AI software overlapping with the PLD’s strict liability regime an explicit clarification should be made that the strict liability regime of the PLD would only apply to High-Risk AI as defined under the AI Act.
• Great care must be taken to ensure the revised PLD’s new requirements will not upset existing national transposition or create new national procedures where no resourcing exists.
• The PLD should maintain its focus on material damages only.
• The PLD’s newer notions on ‘defect’ must be narrowed down to prevent undue justification that can lead to a presumption of a defect and subsequent alleviation of the burden of proof.
• The PLD and AILD must have stronger provisions to prevent the usage of irresponsible Third-Party Litigation Funding and take care against the malicious use of Mass Claims.
• Appropriate safeguards under the PLD and AILD must be created to ensure evidence disclosures are proportionate, so that trade secrets and IP rights are protected. It is paramount that Europe does not introduce ‘discovery clauses’ or promote ‘fishing expeditions’ to the detriment of the defendant.
• The PLD and AILD provisions on the burden of proof de facto establish a situation of ‘guilty until proven innocent’ despite the Commissions efforts to avoid this exact outcome. We do not see sufficient evidence to justify the proposed alleviations.
• Ensure the AI Act is finalised before fully implementing the AI Liability Directive and ensure that AI systems are dealt with under a fault-based liability regime.