The definition of ownership is shifting from physical to algorithmic in the modern age, both in Europe and globally. The legal focus is slowly shifting from simply protecting an idea to demonstrating how that idea was developed. The industry is navigating a landmark collision between traditional copyright and generative AI, alongside a broader redefinition of how technology is shared across borders.

One of the major developments has been the European Parliament’s Resolution on Copyright and AI, published in March 2026. Resolution itself is not a legally binding document – however, it does hold significant value as it determines topics of discussion and sets the legislative agenda for the European Commission. The original 2019 Copyright Directive was more focused on large digital platforms, while the new rules in 2026 will primarily address the training data behind AI. One of the key provisions of the Resolution is that developers can no longer claim that training their models in the US or Asia exempts them from EU law. If an AI tool is accessible to a user in Bratislava or Marseille, the company must comply with the EU copyright transparency, regardless of where the server is located.

Simultaneously, the AI Act’s “sufficiently detailed summary” requirement is likely to be upgraded. Regulators now expect to receive an itemised list of copyrighted works used for model training. This means that creators will be able to see whether their book, song, photo, article, or other artwork was used to build a specific AI model. Lastly, a new proposal for the EU Intellectual Property Office to serve as a central hub for creators to opt out from their work used in AI training purposes. Instead of a photographer notifying dozens of AI companies individually, they could register a single legally binding machine-readable signal at the EUIPO.

The Technology Transfer Reset

From May 2026, the new Technology Transfer Block Exemption Regulation takes effect – a set of rules on know-how sharing that allow companies to enter into licensing agreements without violating EU competition law.

The core concept of TTBER is the so-called safe harbour — a guaranteed safety zone where a licensing agreement of technology rights – such as patents, know-hows, or software copyright – is automatically considered legal and helpful to the economy, and will not be challenged by antitrust regulators. The core concept of TTBER is the so-called safe harbour — a guaranteed safety zone where a licensing agreement of technology rights – such as patents, know-hows, or software copyright – is automatically considered legal and helpful to the economy, and will not be challenged by antitrust regulators, although the “hardcore” violations such as price fixing instantly break that guarantee. Traditionally, when two companies, especially competitors, sign a contract to share technology or patents, competition authorities consider the deal a “cartel,” designed to keep prices high or push competitors out of the market. Proving that a deal is pro-innovation rather than anti-competitive can be a long and expensive legal process.

The EU uses a simple mathematical formula to provide businesses with the needed certainty: If two companies that usually compete with each other share technology, their combined market share must be below 20%; for non-competitors this threshold is 30%. If a business stays within these percentages, there is automatic legal protection. There is no need to proactively prove to the EU that a contract is fair, as the law assumes it is beneficial by default. This allows companies to collaborate and share tech quickly without ongoing an antitrust investigation.

The updated version of TTBER adds a three-year grace period. If a company’s market share exceeds the limit, it will now have three years (instead of two, as in the current rule) to adjust the contracts before losing its safe harbor status. Lastly, the new regulation will include a set of rules regarding data licensing. The European Union is signaling that data represents more than mere information. In the modern age, data is a foundational industrial component, and if a company can’t license the right data, it will struggle to build a final product. Sharing data between companies is usually legally risky as it might resemble information sharing between competitors, which is often problematic under antitrust laws. The new TTBER provides that where firms share the data to produce new goods or services and its market share is under 20% or 30% threshold, the government will not create unnecessary problems, as data licensing falls within the scope of block exemption, provided that it is for the production of contracted goods. This gives companies greater confidence to trade data like any other material.

Standard Essential Patents and Crisis Licensing

The industry expected a game-changing SEP regulation to manage patents inside the technologies such as 5G or Wi-Fi for years. The EU surprisingly abandoned the specific SEP proposal in late 2025. Instead, the year 2026 marks the beginning of a new phase for Union Compulsory Licensing which entered into force on 19 January, allowing the Commission to authorize “Union-wide” use of patented technologies without owner’s consent in cross-bordered crises, such as pandemic or energy shock. This “emergency button” is most interesting addition to the EU legal landscape. The EU concluded that in an emergency, the continent couldn’t afford to let a single patent holder block the production of a vital medicine or a battery component. This creates a strong incentive for patent holders to sign FRAND (fair, reasonable, and non-discriminatory) deals early instead of risking a state-imposed license.

The UCL Regulation is just a specific regime that serves a single purpose, and it is not the SEP regulation the industry was expecting. The new Standard Essential Patent proposal is not expected to start its journey through the legislative procedure during this year, but the business can expect it as soon as 2027.

Operational Readiness in the New Regulatory Landscape

All existing technology transfer agreements should be audited to ensure that they fit the new safe harbour market-share calculations. On 2 August 2026, the transparency obligations under AI Act for GenAI models officially take effect. If a company uses AI to generate content for its clients, it must ensure that it is labeled correctly until that date. By September 2026, the EU is expected to finalise a technical standard for digital watermarking to help the copyright registries identify AI-generated works.

If a company uses AI, it would be advisable to ask the provider for some sort of transparency report. If they can’t provide an itemised list of their training data, the user company may be at risk of contributory copyright infringement under the new rules. At the same time, with the new TTBER, a gap analysis that would help identify which contracts might not fit the safe harbour standards anymore can be of help. Finally, if the company is part of the AV industry, such as advertising or film, the implementation of the machine-readable label for any AI-assisted work is already a must. In 2026, denying the knowledge that the work was generated using AI is no longer a valid legal defence.

Altogether, the current year brings new priorities for the intellectual property and audiovisual industries, and it is important to adapt to the new rules and standards in order to stay competitive and stay out of the danger zone of breaching the EU law.

Read the full Regulatory Horizon Report by B&K Agency: