London image

London

Founded in 2012, Signature Litigation has grown into one of London’s pre-eminent specialist dispute resolution practices. From the outset, we set out to do things differently: a conflict-free, disputes-only platform designed to give clients the undivided focus and strategic agility that complex, high-stakes matters demand.

FIND OUT MORE
Gibraltar image

Gibraltar

Established in 2017, Signature Litigation's Gibraltar office was founded to address growing demand for specialist expertise in commercial litigation and private wealth disputes on the Rock.

FIND OUT MORE
Paris image

Paris

Signature Litigation AARPI houses one of France’s most enviable product liability practices, with the team also handling commercial and corporate litigation, insurance and reinsurance, toxic tort and ESG, civil fraud and asset tracing, international arbitration, administrative and public law.

FIND OUT MORE
Frankfurt image

Frankfurt

Frankfurt has emerged as one of Europe's foremost financial and commercial centres and, increasingly, as a bridge between European and Asian markets. With that growth comes an increasing demand for sophisticated dispute resolution. Signature Litigation established its Frankfurt office to meet this need, bringing our conflict-free, disputes-only platform to the German market.

FIND OUT MORE

Ela Barda comments on the EU AI Act’s transparency obligation in Compliance Week

19 May 2026

Companies “should not treat AI Act compliance in isolation,” said Ela Barda, partner at law firm Signature Litigation. The code expressly states that its transparency obligations sit alongside the EU’s Digital Services Act (DSA), the General Data Protection Regulation (GDPR), the EU Audiovisual Media Services Directive, the European Media Freedom Act, as well as consumer protection and intellectual property (IP) rules. For example, she said, for online platforms already subject to DSA systemic risk assessments (under Articles 34-35), this means integrating AI transparency into existing risk mitigation frameworks rather than building parallel processes.

According to Barda, the code “adds a new compliance layer on top of an already dense regulatory stack”, especially for companies falling both within the scope of “providers” and “deployers”. For example, online platforms sit at the intersection of provider and deployer obligations: they host and distribute third-party AI-generated content, may operate their own generative AI tools, and already face systemic risk obligations under the DSA since very large online platforms (VLOPs) must already use “prominent marking” to ensure that AI-generated or manipulated content resembling real persons, objects, or events is recognizable when presented on their interfaces.

Deployers can avoid disclosure of AI text if a natural or legal person holds editorial responsibility and the content has undergone human review. However, to comply with this, the code requires documented internal procedures identifying the responsible person by name, role, and contact details. For example, said Barda, a platform that permits media publishers to use AI drafting tools and then publish under the publisher’s editorial control may assume the exception applies. “However, if the publisher’s internal documentation is deficient, the exception fails and the platform may face questions about its own role in distributing unlabelled content,” she warned.

The code requires deployers to collaborate on a “best effort basis” with distributors (including platforms) to preserve icons and labels. But in practice, warned Barda, “content is routinely screenshotted, reposted, reformatted, and stripped of metadata as it moves across platforms. Maintaining label integrity across the distribution chain might be technically very difficult.”