As the European Union finalizes the AI Act, a critical question remains unresolved. How should victims seek compensation when AI systems cause harm? The intersection of the Product Liability Directive and AI is central to this debate.
The EU’s Product Liability Directive (PLD) has been the backbone of consumer protection law since 1985. It ensures that companies are held accountable for defective products. But in an era where AI makes autonomous decisions, it evolves over time. AI also operates in a legal gray zone. Lawmakers have questioned whether the PLD is enough.
In September 2022, the European Commission proposed a separate AI Liability Directive. This directive aims to bridge the gap between existing liability rules and the unique risks posed by AI systems. Yet, this proposal has faced intense debate—particularly as the EU AI Act nears completion.
What’s at Stake?
- The EU AI Act sets pre-market rules, but what happens when AI harms someone?
- The AI Act, expected to be finalized soon, regulates AI before harm occurs by classifying AI systems based on risk.
- Nonetheless, it does not set liability rules for when an AI system causes financial, physical, or reputational damage.
- The Product Liability Directive is outdated for AI
- The PLD works well for traditional products (e.g., defective cars, faulty electronics), but it doesn’t cover:
- Standalone AI software that isn’t embedded in a physical product.
- Autonomous decision-making that evolves over time.
- Complex liability chains, where responsibility falls on developers, data providers, or users.
- The PLD works well for traditional products (e.g., defective cars, faulty electronics), but it doesn’t cover:
- The AI Liability Directive would shift the burden of proof
- If passed, the AI Liability Directive would make it easier for victims to sue AI developers. It assumes that AI was responsible unless proven otherwise.
- It would also need AI companies to reveal key internal data to plaintiffs, reducing the “black box” problem.
Industry Resistance: Will This Kill Innovation?
Tech companies and AI developers argue that the AI Liability Directive would stifle innovation by:
- Increasing litigation risk – Companies can be sued for AI mistakes even when they weren’t directly at fault.
- Requiring algorithmic transparency – Some worry this will force AI developers to reveal proprietary trade secrets.
- Duplicating existing laws – Some policymakers believe the PLD can be revised. They think this is preferable to creating an entirely new liability framework.
What The Product Liability Directive and AI Means for Lawyers and AI Compliance
As AI regulation shifts from development to enforcement, attorneys advising clients in AI governance, compliance, and litigation must prepare for:
- Expanded product liability claims – AI-related lawsuits will still fall under existing liability frameworks.
- More legal challenges in AI transparency – The AI Liability Directive will move onward. If it does, expect discovery rules for AI algorithms to change.
- Complex litigation strategies – AI cases will involve multiple parties, evolving software, and difficult causation questions.
🚀 Want to stay ahead of AI liability law? Subscribe to our newsletter at Ethical AI Law Institute for expert analysis on AI litigation and compliance.
📩 Are you advising clients on AI governance? Our AI risk management training for attorneys helps law firms stay compliant with evolving EU regulations.
Leave a Reply