Artificial intelligence and liability: proposed European Commission directives

On 28 September 2022 the European Commission passed two proposals to implement rules on liability to adapt to the digital age, the circular economy and the impact of global supply chains. 

In regard to the first Proposal for a Directive of the European Parliament and of the Council on liability for defective products, it should be noted that this implies repealing the now aging Directive 85/374/EEC (the Product Liability Directive or PLD).

The new Directive lays down rules in common on the liability of economic operators for damage to natural persons caused by defective products.

• Death or personal injury, including medically recognised damage to psychological health.
• Damage to or destruction of property, except: 
  • the defective product itself
  • a product damaged by a defective component of that product
  • property used exclusively for professional purposes.
• Loss or corruption of data that is not used exclusively for professional purposes.

Member States shall ensure that any natural person who suffers damage caused by a defective product (“the injured person”) is entitled to compensation in accordance with the provisions set out in this Directive.

A product shall be considered defective when it does not provide the safety which the public at large is entitled to expect, taking all circumstances into account, including the following:

• the presentation of the product, including the instructions for installation, use and maintenance
• the reasonably foreseeable use and misuse of the product
• the effect on the product of any ability to continue to learn after deployment
• the effect on the product of other products that can reasonably be expected to be used together with the product 
• the moment in time when the product was placed on the market or put into service or, where the manufacturer retains control over the product after that moment, the moment in time when the product left the control of the manufacturer 
• product safety requirements, including safety-relevant cybersecurity requirements
• any intervention by a regulatory authority or by an economic operator referred to in Article 7 relating to product safety 
• the specific expectations of the end-users for whom the product is intended.

Member States shall ensure that national courts are empowered, upon request of an injured person claiming compensation for damage caused by a defective product (“the claimant”) who has presented facts and evidence sufficient to support the plausibility of the claim for compensation, to order the defendant to disclose relevant evidence that is at their disposal.

Member States shall ensure that national courts limit the disclosure of evidence to what is necessary and proportionate to support a claim, and shall consider the legitimate interests of all parties, including third parties concerned, in particular in relation to the protection of confidential information and trade secrets.

Member States shall ensure that, where a defendant is ordered to disclose information that is a trade secret or an alleged trade secret, national courts are empowered, upon a duly reasoned request of a party or on their own initiative, to take the specific measures necessary to preserve the confidentiality of that information when it is used or referred to in the course of the legal proceedings.

• the defectiveness of the product, 
• the damage suffered and 
• the causal link between the defectiveness and the damage.

The defectiveness of the product shall be presumed, when any of the following conditions are met:

• the defendant has failed to comply with an obligation to disclose relevant evidence at their disposal pursuant to what has been stated above; 
• the claimant establishes that the product does not comply with mandatory safety requirements laid down in Union law or national law that are intended to protect against the risk of the damage that has occurred; or 
• the claimant establishes that the damage was caused by an obvious malfunction of the product during normal use or under ordinary circumstances.

The causal link between the defectiveness and the damage shall be presumed, when it has been established that:

• the product is defective and
• the damage caused is of a kind typically consistent with the defect in question.

When a national court judges that the claimant faces excessive difficulties, due to technical or scientific complexity, to prove the defectiveness of the product or the causal link between its defectiveness and the damage, or both, this shall be presumed where the claimant has demonstrated, on the basis of sufficiently relevant evidence, that: 

• the product contributed to the damage; and
• it is likely that the product was defective or that its defectiveness is a likely cause of the damage, or both.

The defendant shall in any event have the right to contest the existence of excessive difficulties or the likelihood referred to above. 

The defendant shall have the right to rebut any of the presumptions (iuris tantum) referred to above.

Member States shall ensure that where two or more economic operators are liable for the same damage pursuant to this Directive, they can be held liable jointly and severally.

Member States shall ensure that a limitation period of three years applies to the initiating of proceedings for claiming compensation for damage falling within the scope of this Directive. 

The limitation period shall begin to run from the day on which the injured person became aware, or should reasonably have become aware, of the following: 

• the damage 
• the defectiveness 
• the identity of the relevant economic operator that can be held liable for the damage.

Member States shall publish, in an easily accessible and electronic format, any final judgment delivered by their national courts in relation to proceedings launched pursuant to this Directive as well as other relevant final judgments on defective product liability.

The publication shall be made without delay upon notification of the full written judgment to the parties. 

The Commission may set up and maintain a publicly available database containing the judgments referred to.

The Commission shall, by six years after the date of entry into force of this Directive, and every five years thereafter, review the application of this Directive and submit a report to the European Parliament, to the Council and to the European Economic and Social Committee.

The purpose of the artificial intelligence (AI) liability Directive is to lay down:

1. Uniform rules for access to information on AI systems.
2. Reduce or ease the burden of proof in relation to damage caused by AI systems and establish broader protection for victims (whether natural persons or companies), which will in turn promote the uptake and rollout of AI by increasing both guarantees and legal certainty for companies.

The Directive will harmonise certain rules for claims outside the scope of application of the Directive on Liability for Defective Products (the Product Liability Directive) in cases where the damage arises due to improper or unlawful conduct (privacy violations or damage caused by a lack of safety).

Just as the fundamental aim of the Artificial Intelligence Regulation is to prevent damage, the AI Liability Directive lays down specific guarantees for victims to obtain reimbursement or compensation when such damage arises.

The Directive simplifies the legal procedure for victims where this concerns proving that somebody’s fault caused damage, by bringing in two core features: 

Firstly: in circumstances where significant fault has been established and it appears reasonably likely that there is a causal link with AI, the so-called “presumption of causality” will embrace the difficulties which victims face in having to explain in detail how the damage was caused by a specific failure or omission, which is something that can be particularly difficult when it comes to understanding and navigating complex AI systems.

Secondly: victims will have more tools with which to obtain legal relief with the introduction of a right of access to evidence (to the relevant proof) in regard to companies and providers for cases where a high-risk AI system is involved.

This Directive will lay down rules in common on: 

• the disclosure of evidence relating to high-risk AI systems to allow claimants to establish grounds for their non-contractual fault-based civil liability claims for damages;
• the burden of proof in the case of non-contractual fault-based civil liability claims filed with national courts for damage and prejudicial consequences caused by AI systems.

This Directive applies to non-contractual fault-based civil liability claims in those cases where the damage and prejudicial consequences caused by an AI-enabled system occur after the end of the transposition period. 

This Directive does not apply to criminal liability.

For the purposes of this Directive, the following definitions will apply inter alia:

“AI system”: an AI system as defined in [Article 3, section 1 of the AI Act].

“High-risk AI system”: a high-risk AI system as defined in [Article 6 the AI Act].

“Provider”: a provider as defined in [Article 3, section 2 of the AI Act].

“User”: a user as defined in [Article 3, section 4 of the AI Act].

“Claim for damages”: a non-contractual fault-based civil liability claim where damages are sought for harm caused by output information from an AI system or for failure of said system to produce output information which it should have generated.

“Claimant”: a person who files a claim for damages and who has been harmed by output information from an AI system or due to the failure of said system to produce output information which it should have generated.

“Potential claimant”: natural or legal person who is weighing the possibility of filing a claim for damages yet has still not done so.

“Defendant”: the person against whom a claim for damages is filed.

“Duty of care”: rule on required conduct laid down under national or EU law to avoid damage to legal assets recognised domestically or by the EU, including life, physical integrity, property and the protection of fundamental rights.

Member states will ensure that national courts are empowered to order the disclosure of evidence, whether (i) at the request of a potential claimant who has previously asked a provider, a person who is subject to the provider’s obligations, or a user to disclose the relevant evidence in their possession with regard to a certain high-risk AI system which is suspected to have caused harm but where their petition has been declined, from such persons, or (ii) at the request of an actual claimant.

In support of this request, the potential claimant must submit sufficient facts and evidence to sustain the feasibility of a claim for damages.

The national courts are to confine the disclosure of evidence and the measures to preserve it to what is necessary and proportionate to support a potential or actual claim for damages.

In deciding whether an order to disclose or preserve evidence is proportionate, national courts will take account of the legitimate interests of all parties, including third parties affected – particularly those interests concerning the protection of trade secrets and confidential information such as, for example, that which relates to public or national security.

When, in the context of a claim for damages, a defendant fails to comply with an order by a national court to disclose or preserve evidence in their possession, the national court will presume non-compliance on the part of the defendant with a relevant duty of care obligation; in particular, in those circumstances to which Article 4, sections 2 or 3 allude, that the evidence requested was intended as proof for the purposes of the associated claim for damages.

The defendant will have the right to rebut this presumption (iuris tantum).

Without detriment to the requirements laid down in said Article 4, national courts will, for the purposes of applying the rules on liability to claims for damages, presume a causal link between the fault of the defendant and the output which the AI system produces or fails to produce, provided that all of the following conditions are met:

• the claimant has demonstrated, or the court has presumed, pursuant to Article 3, section 5 (non-compliance with the relevant duty of care), the fault of the defendant or of a person whose behaviour is the responsibility of the defendant which consists of non-compliance with a duty of care laid down under EU or national law that is directly intended to provide protection against the damage that has been caused;
• it may be considered reasonably likely, based on the circumstances attending each case, that negligent conduct has influenced the output which the AI system produced or failed to produce; 
• the claimant has demonstrated that the information output which the AI system produced, or its failure to produce output, gave rise to the damage.

In the case of claims for damages relating to high-risk AI systems, national courts will not apply a presumption of causality where the defendant proves that the claimant can reasonably access sufficient evidence and expertise to prove the above-mentioned causal link.

In the case of claims for damages relating to non high-risk AI systems, a presumption of causality will only apply where the national court considers it excessively difficult for the claimant to prove the causal link.

In the case of claims for damages against a defendant who has used the AI system during a personal, non-professional activity, the presumption laid down will only apply when the defendant has materially interfered with the conditions of operation of the AI system or when the defendant was required and able to determine the conditions of operation of the AI system and failed to do so. 

Finally, we must point out that the defendant will have the right to rebut the causality presumption.

The Commission shall, within the space of five years after the end of the transposition window at the latest, review application of this Directive and submit a report to the European Parliament, to the Council and to the European Economic and Social Committee accompanied by any proposal of legislation.

The report will examine the effects of Articles 3 (disclosure of evidence and presumption of non-compliance) and 4 (presumption of a causal link) on achieving the objectives pursued by this Directive.

In particular, it must assess the suitability of the rules on strict (no-fault) liability for claims against the operators of certain AI systems – whenever these have not already been regulated by other EU liability rules – and the need for insurance, while at the same time bearing in mind the effect and impact on the general rollout and take-up of AI systems, especially for SMEs.