16 Dec 2024

Who’s to Blame: Unpacking AI Liability Under New EU Rules

Artificial intelligence (AI) is transforming industries and everyday life, but with its rapid development comes new challenges, especially when things go wrong.  The European Union has introduced new rules to make it easier to claim compensation in the event of AI liability.  The new regulations clarify the responsibility of AI system manufacturers and others under certain conditions.  Considering AI systems’ unpredictable nature and development, this is indeed a significant step forward.  It certainly represents a significant progress in fully incorporating AI into everyday life.

Making It Easier to Hold AI Manufacturers Accountable

The newly updated Product Liability Directive aims to cover new technologies, including applications, operating systems, and AI systems.  It will take full effect by December 2026, allowing member states to incorporate the directive into national law.  The new directive specifies that manufacturers who control and are responsible for continued updates of their products are the ones who are liable for damage caused by AI.

This way, individuals no longer need to prove negligence if damage occurs.  They only need to show that the AI product was faulty and caused harm.  This makes it much simpler for people to seek compensation.  However, these rules apply only to individuals.  Businesses that suffer losses caused by AI systems cannot use these regulations to file lawsuits.  However, some EU countries have already imposed corresponding rules for companies.

Manufacturers are the primary liable party.  However, the damaged party can also claim compensation against the importer or the authorized representative established in the EU.  The victim can request information from the distributor of the defective product.  This will help identify the liable party based in the EU.

If no EU-based liable party is identified or the information request is not answered within one month, the victim may seek compensation directly from the distributor.

A Challenge for AI Developers

AI manufacturers and developers are responsible for their products from the moment they introduce them to the market and throughout their entire lifecycle.  Manufacturers will be under increased pressure to regularly track and update their systems to ensure minimal damage risk.  For example, if a company releases an AI program and a problem develops later because it didn’t update or monitor its risks, it can still be liable.  This encourages businesses to take extra care in managing the software and AI systems they sell.

Some tech companies have raised concerns about the new regulations.  They worry that the rules could make innovation more difficult and costly, as businesses may face higher risks when developing AI products.  The interpretation of some parts of the rules, such as what counts as a “faulty” product, remains uncertain.

Serbia’s Regulations on this Matter

Currently, Serbia has no specific AI Act or regulation that addresses AI liability.  However, rules from the Contracts and Torts Act and the Consumer Protection Act can serve as a basis for interpreting a damage claim.   On the other hand, the application of strict liability prescribed by the Contracts and Torts Act (liability for dangerous objects and defective objects) to AI liability is subject to judicial interpretation, depending on the circumstances of the case.

Considering Serbia’s path towards European integration, it will likely follow the EU Directive’s approach and establish liability for AI system manufacturers for damages incurred.  Moreover, it is generally accepted that the most appropriate solution is to hold the manufacturer responsible for damages, with the possibility of extending liability to other parties under certain conditions.

Conclusion

New changes may represent a more significant restriction and responsibility for AI manufacturers.  Regulating liability for damages becomes particularly complex when dealing with something as unique as AI.   The EU’s new rules are a big step toward creating a safer environment for using and developing AI.  We will have to see what the reaction of the state members will be.  With the inevitable development of AI systems and tools, we can always expect new regulations, even beyond the EU level.

 

Author: Marko Jović