Empty Link Skip to Content

The Representative Actions Directive and the Rise of Artificial Intelligence

AUTHORs: Michael Byrne co-author(s): Lorna Murphy Services: Disputes and Investigations DATE: 09/09/2025

The EU Representative Actions Directive

The newly-introduced European Union (“EU”) Representative Actions Directive ((EU) 2020/1828) (the “RAD”) has the aim of enabling consumers within the EU to pool together to protect their interests in a range of areas, such as product liability, data protection, financial services, travel and tourism, energy, and telecommunications. Unlike the position with class actions in the United States, under the RAD, representative actions must be brought by an approved representative entity (called a “Qualified Entity”) on a not-for-profit basis. Qualified Entities can seek injunctive measures, damages, or both, on behalf of the consumers being represented. There are currently 68 pieces of EU legislation which are within the scope of the RAD, meaning that alleged breach of them can form the basis of a representative action. Those pieces of legislation are listed in Annex I of the RAD. In the digital and technology space, they include the General Data Protection Regulation (“GDPR”), the Digital Services Act (“DSA”), the Digital Markets Act (“DSM”), and, from 2 August 2026, the Artificial Intelligence Act (the “AI Act”).

The Representative Actions for the Protection of the Collective Interests of Consumers Act 2023

In Ireland, the transposing legislation for the RAD is the Representative Actions for the Protection of the Collective Interests of Consumers Act 2023 (the “Act”), which was fully commenced on 30 April 2024.  There are now three Irish Qualified Entities capable of bringing representative actions under the Act, namely the Irish Council for Civil Liberties (“ICCL”), NOYB and Digital Rights Ireland. On 26 May 2025, the Irish High Court granted the ICCL permission to proceed with Ireland’s first representative action under the RAD, relating to alleged breaches of the GDPR in connection with online advertising. Looking at the nature of the organisations which have so far applied for Qualified Entity status, it seems likely that data protection and technology will be a particular focus of future claims in Ireland under the RAD.

The rise of US class actions related to Artificial Intelligence

The RAD remains in its infancy, it has not yet been fully implemented in all EU Member States and, to date, only a limited number of claims have been initiated under it across the EU. This being the case, in order to “look into the future” and assess the specific subject matters which might give rise to future claims, it can be instructive to look to other jurisdictions, such as the United States, to evaluate the trends emerging there as regards collective actions. One particular area in which the US is experiencing a growing number of class actions is in the ever-evolving sphere of artificial intelligence (“AI”). 

Biometric privacy / surveillance claims

Multiple collective actions have recently been initiated in the US relating to the use of AI in biometric privacy and surveillance. In March 2025, facial recognition company Clearview AI agreed to pay a settlement sum of $51.75 million to a nationwide class of claimants arising from its use of AI to scrape billions of photographs of individuals’ faces from publicly available webpages.[1] Similar class actions have been taken against Amazon[2] and Home Depot[3] for alleged violation of biometric privacy laws / surveillance.

Misleading advertising claims in relation to AI capabilities

A second category of AI-related class actions gaining momentum in the US are so-called “AI washing” claims brought against companies for alleged overstatement of their AI progress or features.  Numerous class actions have been initiated in the US in recent years alleging misrepresentation and securities fraud against companies for allegedly falsely exaggerating their AI capabilities.  Zillow[4], Innodata[5], GitLab[6], Oddity Tech[7] and Evolv Technologies[8] are just some of the companies which have recently faced this type of class action.

AI-related copyright claims

A significant number of AI-related class action claims have also been taken in the US by authors alleging non-consensual use of their content to train AI models. The largest copyright class action ever certified was filed in 2024 in California by three book authors who claim that AI operator Anthropic used their copyrighted works to train its AI-powered chatbot, Claude.[9] It had been reported that the class action against Anthropic could potentially be joined by up to seven million plaintiffs. According to recent media reports, Anthropic has agreed in principle, subject to the approval of the Court, to settle the case for a sum of at least $1.5 billion. This figure is based on a payment of $3,000 per affected work and it is estimated that there are at least 500,000 works involved.  

As AI continues to shape industries and impact everyday decision-making, legal systems are being tested on how to respond to the novel forms of alleged harm that can occur at scale.  While AI is here to stay, companies should be cognisant of the litigation risks associated with its use, and recent class actions in the US may serve as a warning in this regard.  We will have to wait to see if similar trends repeat themselves in the Irish and European courts in the future, and whether the RAD will prove to be a mechanism for litigating these novel types of claims.

Contact Us

If you have any queries about anything discussed in this insight, please contact Disputes and Investigations partner Michael Byrne, or your usual Matheson contact.

References

[1] In re Clearview AI, Inc. Consumer Privacy Litigation, No. 1:21-cv-00135 (the United States District Court for the Northern District of Illinois)
[2] Kaeli Garner, et al v Amazon.com Inc, et al., No.2:21-cv-00750-RSL
[3] Benjamin Jankowski v The Home Depot, No. 1:25-cv-09144
[4] Dibakar Barua, et al. v Zillow Group Inc. et al., No. 2:21-cv-01551-TSZ
[5] David D’Agostino v Innodata Inc., Jack Abuhoff, Mark Spelker, and Marissa Espineli, No. 2:24-cv-00971
[6] Dolly v. GitLab Inc., No. 24-cv-06244 (N.D. Cal.)
[7] Hoare v. Oddity Tech Ltd. et al., No. 1:24-cv-05037
[8] Raby v. Evolv Technologies Holdings, Inc. f/k/a NewHold Investment Corp. et al>No. 1:24-cv-10761
[9] Bartz v. Anthropic PBC, U.S. District Court for the Northern District of California, No. 3:24-cv-05417