Artificial Intelligence and Intellectual Property: Continuity, Transformation and Protection Challenges
Introduction
Over the past sixty years, technological progress has profoundly transformed both the subject matter and the practice of intellectual property. From industrial innovations to advances in life sciences and digital technologies, each wave of innovation has raised new questions regarding protection and required existing legal frameworks to evolve, while preserving the fundamental principles at the core of IP law.
Artificial intelligence is part of this ongoing evolution, while introducing challenges of its own. Through its rapid development and wide range of applications, AI has become an essential technology for businesses, offering significant opportunities in terms of efficiency, innovation and decision support, while also creating new uncertainties relating to rights, confidentiality and control over information.
These developments directly affect intellectual property, a strategic asset for companies but one whose interaction with AI remains insufficiently understood in many cases. They also raise important questions for the practice of intellectual property professionals such as Patent and trademark attorneys and IP lawyers), whose role is precisely to analyze technical innovations and design appropriate protection strategies in constantly evolving technological environments.
This article examines:
- the integration of AI into everyday business activities and the legal issues it raises;
- its impact on the profession of intellectual property consultants;
- and the challenges associated with protecting AI-related technologies.
1. AI in Everyday Use: Rights, Confidentiality and Control of Information
Generative AI is now widely used for content creation, assistance with numerous tasks and technical development activities. Rapidly adopted by businesses, its use raises significant legal risks, particularly in relation to intellectual property and information management.
In this context, a structured governance approach is essential. This includes implementing internal policies and appropriate training programs to raise awareness of:
- the opportunities and limitations of AI tools;
- the risks associated with the use of confidential information;
- and the importance of respecting third-party rights.
Such measures help ensure that AI is used responsibly, balancing its operational benefits with applicable legal requirements.
Risks Relating to Confidentiality and Trade Secrets
The use of AI tools requires the input of data, which may include sensitive information. Depending on the terms of use, this information may be stored, reused or exploited for model training purposes.
This creates a risk of loss of confidentiality, particularly with regard to trade secrets. In the field of industrial property, such disclosure may also jeopardize the patentability of an invention, especially where novelty requirements are concerned.
Particular attention should therefore be paid to the contractual terms, technical implementation and security features of the tools being used, as well as to the nature of the information provided to them.
Risks of Infringing Intellectual Property Rights
AI systems are typically trained on vast quantities of data, including content protected by intellectual property rights. This raises questions regarding the lawfulness of training data, the operation of the models themselves and the legal status of generated outputs.
In certain circumstances, AI-generated outputs may reproduce or imitate protected elements, exposing users to potential infringement risks, particularly in relation to copyright and trademarks.
A specific issue also arises with AI-generated software. Generated code may unknowingly incorporate elements governed by open-source licenses. Such licenses can impose restrictive obligations, including disclosure requirements or attribution obligations, which may conflict with a company’s intended protection or commercialization strategy.
Uncertainties Surrounding AI-Generated Content
Beyond infringement concerns, AI also raises questions regarding the nature and legal status of generated content.
On the one hand, outputs may be inaccurate, inconsistent or unsuitable for their intended purpose. Given the complexity of these models, it remains difficult to predict their behavior reliably across all situations.
On the other hand, ownership and protection of AI-generated content remain uncertain. In many jurisdictions, particularly in Europe, copyright protection is based on the existence of human authorship. As a result, content generated autonomously by AI may not qualify for protection, or its legal treatment may depend heavily on the degree of human involvement.
Furthermore, the contractual terms governing AI tools often contain specific provisions regarding the use and exploitation of generated outputs, adding another layer of complexity.
These uncertainties call for a cautious approach, especially when AI-generated content plays a strategic role within a business.
2. The Impact on the Intellectual Property Profession
AI directly affects intellectual property professionals, particularly through the emergence of generative tools that are especially effective in producing written content.
Emerging Uses: Between Automation and Assistance
At present, the most realistic applications of AI in professional practice do not involve full automation but rather targeted assistance. AI is particularly useful for automating repetitive tasks such as formatting, document structuring and data management, for developing internal tools and agents, and for supporting drafting and analytical work.
Professional AI solutions nevertheless share a fundamental limitation: generative models are probabilistic by nature. They produce plausible outputs, but without any guarantee of accuracy. Unlike deterministic tools, they can generate inconsistencies, errors or even hallucinations. Their use without critical review may therefore create significant legal risks.
As with businesses generally, the use of these tools also raises ongoing confidentiality concerns, depending on their architecture and terms of use.
Moreover, the growing reliance on “one-click” solutions may weaken professional engagement. The rapid generation of polished text can encourage superficial validation at the expense of strategic thinking and careful analysis.
In practice, these tools often produce content that is statistically plausible but strategically average: patent applications that may appear satisfactory at first glance but lack the differentiation and optimisation required for robust protection.
For these reasons, AI cannot be regarded as a reliable tool in the same way as a calculator. It requires critical oversight and careful use by IP practitioners.
Strengthening the Quality and Consistency of IP Analysis
Despite these limitations, the integration of AI into professional practice appears inevitable. When used appropriately, it can be a valuable tool for:
- facilitating document review and synthesis;
- strengthening the quality and consistency of patent drafting and other written work;
- testing alternative wording or legal arguments;
- and challenging an analysis through simulated objections.
In our view, this is the most appropriate approach today: using AI as a tool for reflection and discussion rather than as a substitute for professional judgement.
AI can provide a form of “map” of possible solutions by identifying alternative formulations or analytical approaches. However, it cannot:
- fully understand a client’s context;
- assess economic, commercial or litigation-related constraints;
- or assume responsibility for strategic decisions.
These remain matters of professional expertise. The intellectual property consultant remains the only person capable of determining the most appropriate protection strategy.
Ultimately, the principle remains simple: AI assists, humans decide.
3. Protecting AI Through Patents
The rise of AI has also led to a significant increase in patent filings, particularly before the European Patent Office (EPO).
From the EPO’s perspective, AI-related inventions are generally treated as computer-implemented inventions.
As such, they remain subject to the established principles of the European Patent Convention, notably:
- the exclusion of mathematical methods and computer programs “as such”;
- and the requirement that an invention possess technical character in order to be patentable.
This approach is firmly rooted in established case law, particularly the COMVIK decision (T 641/00), according to which only features contributing to technical character can support inventive step.
Contrary to a common misconception, AI does not constitute a separate legal category. It falls within a legal framework that has long existed for software-related inventions.
Identifying a Technical Effect
The particularity of AI lies in its underlying nature: it is based on mathematical models and algorithms, which are in principle excluded from patentability as such.
The central issue is therefore the transition from the abstract to the technical.
To be patentable, an AI-related invention must demonstrate either:
- a technical effect resulting from its application (for example image processing, signal detection or the control of a technical system);
- or a technical contribution linked to its implementation (for example resource optimisation or improved operation of a computer system).
The EPO therefore accepts the patentability of many AI-based applications when they are integrated into a technical context. For example, the use of a neural network to detect anomalies in medical data or to process signals will typically be regarded as providing a technical contribution.
By contrast, AI used solely for abstract purposes, such as purely conceptual classification or data processing without a technical purpose, will generally remain excluded.
Although the distinction is conceptually well established, its practical application is often complex and requires careful drafting and a rigorous demonstration of the technical effect achieved.
Increased Requirements for Drafting and Protection Strategy
The specific characteristics of AI require particular care when preparing patent applications. It is often necessary to:
- clearly define the technical problem being solved;
- explicitly describe the technical effects achieved;
- and, where relevant, characterize aspects of the model or training data that contribute to those technical effects.
At the same time, the complexity and sometimes opaque nature of AI systems can make it difficult to strike the right balance between sufficient disclosure and the preservation of a competitive advantage.
Protecting AI-related innovations through patents therefore requires a careful analysis of the technical effects actually achieved — a task that lies at the heart of the patent practitioner’s expertise.
Conclusion
Artificial intelligence has become an essential technology, rapidly adopted by businesses and capable of generating significant opportunities. At the same time, it requires increased vigilance, both in its use and in the protection of the innovations it enables.
For intellectual property professionals, AI represents neither a complete disruption nor merely another technological development. Rather, it forms part of a broader process of continuous adaptation to technological change, which has always been a defining characteristic of the profession.
In this respect, current developments do not fundamentally redefine the profession; they reinforce its core requirements. Analysis, judgement and responsibility remain central to the role of the IP practitioner in an environment where tools evolve rapidly but protection challenges remain critical.
As Office Freylinger celebrates its sixtieth anniversary, this continuity becomes particularly evident: supporting innovation, understanding its technical implications and securing appropriate protection for businesses, regardless of how technology evolves.


