How the EU AI Act Will Change IP Enforcement and Litigation in Greece, Practical Steps for Rights‑holders

The intersection of AI and IP in Greece has entered a critical new phase. With the EU Artificial Intelligence Act (Regulation (EU) 2024/1689) now fully applicable and its enforcement architecture taking shape across Member States, rights‑holders operating in or targeting the Greek market face a fundamentally altered litigation landscape. Greek IP law, anchored in Law 2121/1993 on copyright and Law 4072/2012 on trademarks, has not been repealed, but the procedural tools available to claimants, the evidence they can demand, and the parties they can pursue have all expanded.

This guide provides an operational playbook for in‑house counsel, IP owners, and platform operators who need to understand how the EU AI Act changes IP enforcement across borders, and what concrete steps to take in Greece right now.

Executive Summary and Quick Actions for Rights‑Holders in AI and IP Greece Disputes

Before diving into the legal detail, here are the immediate tactical steps every rights‑holder should take to protect IP assets in an environment shaped by generative AI and the EU AI Act in Greece.

Quick Checklist, Top Actions to Take Now

• Audit your IP portfolio for AI exposure. Identify which copyrighted works, trademarks, or designs are most likely to appear in AI training datasets or generative outputs. Prioritise high‑value assets.
• Preserve evidence of AI‑generated infringement immediately. Screenshot outputs, record timestamps, capture URLs, and use notarised screen captures (a well‑established Greek evidentiary mechanism) before content is altered or removed.
• Send a formal preservation notice to the AI provider. Cite Article 16 of the EU AI Act (provider obligations on technical documentation and logging) and request that all relevant model logs, training data records, and output histories be retained.
• Issue a takedown notice to the hosting platform. Under the Digital Services Act (Regulation (EU) 2022/2065) and Greek implementing provisions, platforms must act expeditiously on valid IP notifications. Document the notice and response timeline.
• Review and update your licensing agreements. Add explicit AI‑related clauses covering machine‑learning ingestion, model training, and synthetic output rights. Contracts signed before mid‑2024 almost certainly lack these provisions.
• Map the liability chain. Identify the provider (model developer), deployer (the business using the model), and any intermediary platform. Each bears different obligations under the AI Act, and your enforcement strategy must target the right entity.
• Engage Greek litigation counsel early. Greek interlocutory injunction procedures (Articles 682–703 of the Greek Code of Civil Procedure) can deliver rapid provisional relief, but the evidentiary thresholds require careful preparation.
• File a regulatory complaint where appropriate. Article 77 of the EU AI Act requires Greece to designate national competent authorities for market surveillance. A parallel regulatory track can reinforce private enforcement.

Three Immediate Templates

Rights‑holders in AI and IP Greece disputes should prepare these documents without delay:

• Evidence Preservation Letter. Addressed to the AI provider, citing Article 16 and Article 50 of the EU AI Act, requesting retention of model logs, training data provenance records, and output histories linked to the allegedly infringing content.
• Platform Takedown Notice. Structured to meet the Digital Services Act Article 16 requirements: clear identification of the infringing content, the IP right at issue, the location of the content, and the basis for the claim.
• Internal AI‑Audit Memo. A short internal document cataloguing which company IP assets are exposed, what generative AI tools employees or contractors use, and where licensing gaps exist.

What the EU AI Act and Greek AI Laws Require, The Legal Baseline

Understanding the regulatory architecture is essential before launching any IP enforcement action. The EU AI Act does not directly regulate intellectual property rights, but its obligations on providers, deployers, and distributors create new disclosure duties, documentary trails, and enforcement leverage that IP litigators in Greece can exploit.

Key Provisions of the EU AI Act Relevant to IP Enforcement

• Article 16, Provider obligations. Providers of high‑risk AI systems must maintain technical documentation, implement quality management systems, keep logs generated by their systems, and ensure traceability. For IP litigators, these records can serve as crucial evidence of what training data was used and how outputs were generated.
• Article 50, AI transparency obligations. Providers of AI systems that generate synthetic content (text, images, audio, video) must ensure that outputs are marked in a machine‑readable format as artificially generated or manipulated. Deployers must also disclose AI involvement in certain contexts. This transparency obligation makes it harder for infringers to conceal AI‑generated content and creates a documentary trail rights‑holders can follow.
• Article 53, Obligations for general‑purpose AI models. Providers of general‑purpose AI (GPAI) models must draw up and maintain technical documentation, including a sufficiently detailed summary of the content used for training. This is the provision most directly relevant to claims that an AI model was trained on copyrighted material.
• Article 77, National competent authorities. Each Member State must designate at least one notifying authority and at least one market surveillance authority to oversee AI Act compliance. Greece is required to establish or designate these authorities, creating a regulatory channel that rights‑holders can use alongside private litigation.

Greek National Framework, Law 4961/2022 and Related Legislation

Greece adopted Law 4961/2022 (“Emerging Information and Communication Technologies, Strengthening of Digital Governance”), which established a national framework for emerging technologies including AI. The law created the institutional basis for AI governance in Greece, assigning oversight functions and promoting the responsible development of AI systems. While Law 4961/2022 predates the EU AI Act and does not replicate its risk‑classification system, it provides the institutional infrastructure, including the role of the Ministry of Digital Governance, through which Greece is expected to implement and enforce AI Act obligations.

For data protection, Law 4624/2019 transposed the GDPR into Greek law and established the Hellenic Data Protection Authority (HDPA) as the supervisory body. This is directly relevant to AI and IP enforcement because requests for training data, model logs, and user information will frequently intersect with personal data protections.

Timeline, EU AI Act and Greece Implementation Milestones

How Liability for AI Outputs Is Allocated, Practical Mapping

One of the most pressing questions in AI and IP Greece disputes is: who do you sue? When a generative AI system produces content that infringes a trademark, reproduces copyrighted material, or imitates a protected design, the liability for AI outputs must be traced through the entire chain, from the model developer to the end‑user.

Scenario 1, AI Model Trained on Copyrighted Text

A Greek publisher discovers that a GPAI model reproduces substantial portions of its copyrighted articles in response to user prompts. The question of ai copyright greece is not resolved by the AI Act itself, copyright remains governed by Greek Law 2121/1993 and the EU copyright acquis (including the DSM Directive 2019/790). However, the AI Act strengthens the publisher’s hand: under Article 53, the GPAI provider must maintain a training data summary. The publisher can demand this summary, cross‑reference it with its own catalogue, and use the results as evidence of unauthorised reproduction or adaptation. If the provider failed to respect an opt‑out under Article 4 of the DSM Directive, the liability case is substantially reinforced.

Scenario 2, Platform Re‑Serving AI‑Generated Output That Infringes a Trademark

A luxury brand discovers that a Greek e‑commerce platform hosts product listings with AI‑generated images that closely mimic its registered trademarks. In this scenario, the enforcement strategy bifurcates: the platform receives a takedown notice under the DSA framework, while the deployer (the seller using the AI tool to generate listings) faces a direct infringement claim. The AI provider may also be pursued if the model was specifically designed or marketed for creating product imagery without adequate safeguards against trademark reproduction. IP enforcement against generative AI outputs requires identifying and pursuing all three links in the chain simultaneously.

Evidence and Disclosure Under the AI Act, What to Demand and How to Use It in Greek Courts

The evidentiary landscape is where the EU AI Act’s impact on IP litigation in Greece may be most transformative. Before the Act, rights‑holders had limited tools to compel disclosure of how an AI model was trained, what data it ingested, or how it generated a particular output. The AI Act’s documentation and logging obligations, particularly Articles 16, 50, and 53, change this fundamentally.

Preserving Model Outputs and Logs, AI Infringement Evidence

Greek courts accept a broad range of evidence in civil proceedings, including electronic documents, notarised screenshots, expert reports, and official records. For AI‑related IP disputes, the critical evidence categories are:

• Output capture. Dated, notarised screenshots or screen recordings of the infringing AI output, including the prompt used (if visible), the platform or application, and the URL or access point.
• Model logs. Under Article 16 of the AI Act, providers of high‑risk AI systems must keep logs automatically generated by their systems. Even for GPAI models not classified as high‑risk, Article 53 requires technical documentation. Rights‑holders should request these logs immediately upon identifying infringement.
• Training data summaries. Article 53 requires GPAI providers to prepare a sufficiently detailed summary of training data. This is the single most powerful new disclosure tool for IP rights‑holders: it can reveal whether copyrighted works were included in training sets.
• Metadata and watermarking. Article 50 mandates that AI‑generated content be marked in a machine‑readable format. Failure to comply with this obligation is itself evidence of a regulatory breach that can strengthen an infringement claim.

Draft Disclosure Request Language

When sending a pre‑action disclosure request to an AI provider in connection with Greek IP proceedings, the letter should include the following elements:

• Identification of the IP rights at issue (registration numbers, dates, scope of protection)
• Description of the allegedly infringing output with supporting evidence (screenshots, dates, URLs)
• A specific request for: (a) training data summary prepared under Article 53 of the EU AI Act; (b) automated logs retained under Article 16; (c) any transparency markings or metadata applied under Article 50; and (d) records of the deployer or user who generated the output
• Legal basis for the request: the AI Act obligations, combined with the rights‑holder’s entitlement to information under Article 8 of Directive 2004/48/EC (the IP Enforcement Directive), as implemented in Greek law
• A reasonable deadline for response, typically 14–21 days, with a notice that failure to respond will be cited in any subsequent Greek court proceedings

Using Administrative and Regulatory Channels

Under Article 77 of the EU AI Act, national competent authorities have investigation and enforcement powers. Once Greece formally designates its market surveillance authority for AI, rights‑holders will be able to file complaints requesting investigations into non‑compliant AI systems. A parallel regulatory complaint can serve multiple purposes: it may compel disclosure that would be difficult to obtain through private litigation alone, it creates an official record of non‑compliance that supports civil claims, and it may result in administrative orders that require the provider to halt or modify the infringing system.

Industry observers expect that the Greek market surveillance authority, once operational, will coordinate closely with the Hellenic Copyright Organisation (OPI) and the Hellenic Industrial Property Organisation (OBI) on IP‑related AI complaints.

Injunctions, Takedowns and Provisional Remedies, Greek Practice and Tactics

Greek law provides a robust framework for provisional remedies that is well‑suited to the speed and scale of AI‑related IP infringement. The interlocutory injunction procedure under Articles 682–703 of the Greek Code of Civil Procedure allows rights‑holders to obtain rapid interim relief before a full trial, making it a critical weapon in ip enforcement generative ai disputes.

Injunction Checklist, Key Requirements in Greek Courts

• Urgency. The applicant must demonstrate an imminent danger or urgent need that justifies provisional relief before a main action is heard. With AI systems capable of generating thousands of infringing outputs per hour, establishing urgency is typically straightforward.
• Prima facie case. The court must be satisfied that the applicant has a plausible claim of IP infringement. This requires presenting the registered right (or evidence of unregistered protection), the infringing output, and a credible chain linking the output to the respondent.
• Proportionality. The requested relief must be proportionate to the harm. Courts will weigh the burden on the respondent, particularly if the injunction would require shutting down an entire AI service, against the harm to the rights‑holder.
• Security for costs. The court may require the applicant to post a bond or security to cover the respondent’s losses if the injunction is later found to have been unjustified.

Tactical Considerations, Provider vs Platform

When the infringer is an AI provider based outside Greece, obtaining and enforcing an interlocutory injunction raises cross‑border enforcement challenges. The Brussels I Recast Regulation (Regulation (EU) 1215/2012) governs jurisdiction and recognition of judgments across EU Member States, allowing a Greek injunction to be enforced in the provider’s home jurisdiction without an exequatur procedure. Where the provider is outside the EU, rights‑holders should consider pursuing the hosting platform within Greece simultaneously, as platforms with Greek users and Greek‑language services are subject to Greek jurisdiction under both the DSA and general private international law principles.

For evidence bundles supporting an emergency injunction application, rights‑holders should prepare: notarised output captures, an expert affidavit explaining how the AI system produced the infringing content, the formal preservation notice sent to the provider, proof of the registered IP right, and any response (or non‑response) from the provider or platform.

Damages, Remedies and Valuation Where AI Is Involved

Quantifying damages in AI‑related IP disputes requires adapting traditional valuation models to account for the scale, speed, and often untraceable distribution of AI‑generated infringing content. Greek courts apply the general damages framework under Law 2121/1993 (for copyright) and the IP Enforcement Directive (2004/48/EC), which offers three calculation methods:

• Lost profits. What the rights‑holder would have earned absent the infringement. In AI contexts, this may include lost licensing revenue from datasets or content that was ingested without authorisation.
• Unjust enrichment. The profit the infringer obtained by using the IP without permission. Where a deployer used an AI system to generate commercial content based on copyrighted material, the revenue generated from that content is relevant.
• Reasonable royalty. The licence fee that would have been agreed between a willing licensor and licensee. This model is particularly useful where the rights‑holder would have been willing to license AI training rights at a market rate.

Expert Evidence, Economists and Forensic AI Specialists

Greek courts rely on expert evidence for complex damages calculations. In AI‑related IP cases, two types of expert are typically needed: an economist to model the financial loss using one of the three methods above, and a forensic AI specialist to establish the technical link between the training data, the model, and the infringing output. Early indications suggest that Greek courts will increasingly appoint or accept court‑appointed experts with dual expertise in data science and IP valuation, reflecting the technical complexity of these disputes.

Data Protection and Trade Secrets, Limits on Evidence Collection

Every request for training data, model logs, or user information in an AI and IP Greece dispute will intersect with data protection law and trade secret protections. Rights‑holders must navigate these constraints carefully to avoid having evidence excluded or claims dismissed.

GDPR and Law 4624/2019 Constraints

Law 4624/2019, Greece’s GDPR implementing statute, governs the processing of personal data by both public and private entities. When requesting training data or output logs from an AI provider, the rights‑holder must consider whether the data includes personal information, as it frequently will, particularly where the AI was trained on publicly available internet data. Requests should be framed as narrowly as possible, targeting metadata, content identifiers, and technical logs rather than raw personal data. Where personal data disclosure is unavoidable, the requesting party should propose redaction protocols and, where possible, obtain a court order specifically authorising the disclosure under the procedural safeguards of the Greek Code of Civil Procedure.

Trade Secret Protections

AI providers will frequently assert that model weights, training methodologies, and detailed dataset compositions constitute trade secrets protected under the EU Trade Secrets Directive (2016/943) and Greek implementing legislation. Greek courts can and do impose confidentiality orders limiting who may access disclosed material. The likely practical effect is that rights‑holders will receive training data summaries rather than complete datasets, and model documentation rather than source code, but even these redacted disclosures, mandated by the AI Act, represent a significant evidentiary advance over the pre‑Act landscape.

Practical Checklist for Rights‑Holders, Preserve, Prove, Plead

The following step‑by‑step checklist consolidates the operational guidance in this article into a single printable reference for rights holders in AI Greece enforcement actions.

Pre‑Action Timeline Template

• Day 1–3: Preserve. Capture all infringing outputs with notarised screenshots. Send preservation notices to the AI provider and hosting platform. Instruct external forensic experts to archive outputs and metadata.
• Day 4–7: Identify. Map the liability chain, provider, deployer, platform. Confirm IP registrations and gather registration certificates. Check whether the AI system is classified as high‑risk under the EU AI Act.
• Day 7–14: Demand. Send the formal disclosure request letter (citing Articles 16, 50, and 53 of the AI Act and Article 8 of the IP Enforcement Directive). Set a 14–21 day response deadline.
• Day 14–21: Assess. Review provider response. Instruct economist and forensic AI experts to prepare preliminary damages assessment and technical linkage report.
• Day 21–30: Plead. If no satisfactory response, file an interlocutory injunction application under Articles 682–703 of the Greek Code of Civil Procedure. Simultaneously file a regulatory complaint with the competent national authority. Prepare the main action for filing.
• Ongoing: Monitor. Continue monitoring for new infringing outputs. Update evidence bundles. Review and update licensing agreements to include AI‑specific clauses.

Connecting AI and IP Greece Strategy to Broader Enforcement

AI‑related IP enforcement in Greece does not operate in isolation. Rights‑holders with international portfolios should coordinate Greek proceedings with parallel actions in other EU jurisdictions, leveraging the Brussels I Recast framework for cross‑border recognition and enforcement. The international intellectual property guide on this site provides comparative context for multi‑jurisdictional strategies. Similarly, rights‑holders dealing with broader legal changes in Greece for 2026 should factor evolving property and regulatory law into their overall risk assessments.

The convergence of the EU AI Act, Greek national legislation, and established IP enforcement mechanisms creates both challenges and powerful new opportunities for rights holders in AI Greece disputes. Those who act early, preserving evidence, mapping liability chains, and preparing disclosure requests grounded in AI Act obligations, will hold a decisive strategic advantage as this area of law matures.

Sources

1. Regulations.ai, Greece AI Regulation Summary
2. Zepos & Yannopoulos, AI Regulation in Greece
3. PwC Greece, Legal Flash: EU AI Act
4. Chambers Practice Guides, AI Regulation Comparison
5. DLA Piper Intelligence, AI Laws of the World: Greece
6. Greek Customs (AADE), IPR Customs Enforcement
7. CEE Legal Matters, Greece: EU AI Act Key Steps for Enterprises