They can be your spearhead and shield, driving the exclusivity of each solution, building customer trust, and attracting investment by securing a strong position in niche markets. An effective IP strategy can often involve making multiple decisions rather than choosing between alternatives. IP rights are typically categorized into three groups: copyright, industrial property (such as trademarks, patents, and utility models), and "soft IP" (which includes trade secrets, know-how, and confidentiality). Owning an IP right grants temporary control over your creation, preventing unauthorized use or commercialization.
For AI-based software solutions, Data protection is paramountl. To support businesses, Europe has introduced the world's first comprehensive legal framework on AI, the “AI ACT”In force since August 1, 2024, this regulation, while not directly addressing intellectual property rights, supports compliance with the EU Copyright Directive (Directive 790/2019) regarding the scraping of databases for AI training by improving security measures. Currently, data mining for scientific purposes is permitted without the need for authorization from the rights holder, but broader extraction requires explicit permission from the owners. In all cases, legitimate access to the data is essential.
Copyright remains by far the most common method for software in the EU, protecting databases as compilations and defending the companies that own them and the substantial investment in their content. In cases involving AI-generated data sets or data processed for training, copyright protection extends to the selection or arrangement (but not to the content itself). This approach aligns with the EU Database Directive (EU's Database Directive), which provides protection for databases that demonstrate a substantial investment in obtaining, verifying or presenting data. European Commission has also introduced several data-related regulations, including the Data Governance Act (Data Governance Act), to harmonize ownership concepts across member states.
Finally, in some cases, patents for AI solutions may also be applicable and effective. Let's look at these two types of protections in detail:
About Copyright©️: The software underlying AI models may be protected by copyright, but raw data lacks the originality to be protected. Instead, collections of data may be granted database rights if a significant investment was made in their organization, although this excludes machine-generated data. Notably, copyright lasts for 70 years after the author's death.
However, copyright protection is limited in scope as it focuses on the creative organization and presentation of data but does not secure the raw inputs used by AI models. As a complement to this protection, contractual agreements become vitally important. Contracts can clearly delineate data ownership, usage rights, and exploitation terms, providing businesses with an additional layer of protection where existing IP laws may be insufficient. For example, defining the ownership of machine-generated data through agreements can mitigate potential disputes over control and commercial exploitation of AI-derived knowledge.
About Patents®️: At European level, applicants for protection of AI innovations through patents must demonstrate that their invention contributes a novel technical solution to an existing problem. These technical contributions may include advances in image recognition, speech processing or medical diagnostics – domains where AI-driven processes significantly improve efficiency or accuracy. A successful AI patent application is based on the problem-solution approach, ensuring that the patent claims show technical contributions, such as new data processing techniques, advanced model architectures or new ways of interacting with hardware.
It is important that AI patent claims focus on data generation and collection, unique model training methodologies, and the technical applications of these models. This involves going beyond abstract concepts, which are not patentable, and ensuring a robust framework of human interaction and oversight within the innovation. When filing the patent, the biggest challenge lies in ensuring that the claims accurately reflect the technical contributions without overly broad or generic descriptions, which could jeopardize validity during examination or litigation.
It is also essential to consider that patents remain valid for 20 years from the date of application, after which they must be re-evaluated for their continued relevance and value in a rapidly evolving technological landscape.
Adapting to evolving frameworks and international inconsistencies
The evolving nature of AI challenges traditional legal frameworks globally. Issues such as AI invention and data ownership are at the forefront of ongoing dialogues as new AI technologies continue to push the boundaries of existing legal concepts. For example, Europe and the United States Both countries maintain that only humans can be considered inventors, a principle protected in the legal frameworks of both regions. However, in a landmark case, South Africa accepted an AI system as an inventor, a decision that raises questions about the future direction of patent laws globally. Companies navigating this rapidly changing landscape must develop comprehensive IP strategies that leverage multiple layers of protection. For example, combining technical patent protection with strong database rights, confidentiality agreements, and strategic trade secret policies can offer a multifaceted defense.
In this context, establishing a strong and agile IP portfolio not only protects companies against infringement, but also serves as a powerful spearhead for market positioning, licensing negotiations and competitive advantage. Drive your vision with strategy and determination!
EU priorities: generative AI
The EIC Accelerator has announced a specific challenge for Artificial Intelligence (AI) in 2025 (IV.2.3 GenAI4EU: Creating European Champions in Generative AI), reinforcing the EU's commitment to strategic technological leadership. As AI continues to drive transformation in key sectors such as industry, healthcare, sustainability and security, the EU has decided to prioritise investment in high-impact, market-ready AI innovations. This initiative seeks to support SMEs developing disruptive AI solutions, aligning with the European agenda to strengthen digital sovereignty, economic resilience and global competitiveness in this field.
Flavio Nucciarelli,
Euro-Funding Project Manager
Any questions about this news?
Talk to Coexia, the AI of the foreign trade
