Understanding AI Liability: Navigating the Legal Landscape

← All articles
Artificial intelligence · EN

Understanding AI Liability: Navigating the Legal Landscape

AI’s potential for businesses is undeniable, but the question of liability for AI-caused damages is a complex issue. Generally, one is liable for intentional and negligent actions, or if contractual agreements are not met. This applies to AI as well, with the user, developer, or operator generally being liable, provided they adhere to laws and contractual agreements.

However, the AI landscape in the EU is filled with norms imposing duties of care and contractual requirements, including the Product Liability Act, Civil Code and General Terms and Conditions.

New EU regulations aim to make this legal field clearer and more adapted to future developments:

- Revised Product Liability Directive,
- AI Act,
- AI Liability Directive (attached).

The revision of the Product Liability Directive now considers software a product and extends liability to the manufacturer, importer, authorized representative of the manufacturer, fulfillment service provider, retailer, and operator of online marketplaces. This means that even for AI products purchased outside the EU, there is a claimant who has to compensate for damage.

The AI ACT defines who is a provider and what an AI system is, differentiating between high-risk AI and other AI systems. This classification has significant implications for contract design when acquiring AI systems.

The AI Liability Directive applies when damage has occurred through or during the operation of an AI system. It complements civil liability in relation to intentional and negligent behavior, regardless of contractual liability.

Both the revised Product Liability Directive and the AI Liability Directive apply concurrently, with aspects of the AI Liability Directive incorporated into the revised Product Liability Directive. However, navigating these regulations can be challenging due to the need to refer to the AI ACT, the revised Product Liability Directive, and the AI Liability Directive.

Before implementing AI, it is vital to create contractual foundations, implement suitable technical and organizational measures, and incorporate data protection into the evaluation. The future will undoubtedly bring changes, and it is crucial to stay informed and prepared.

#AILiability #AI

This post has a document attached on LinkedIn. View on LinkedIn →

Discuss on LinkedIn