The use of AI tools such as ChatGPT, translation systems or code assistants has already become part of everyday life in many companies. In many cases, however, this takes place without clear guidelines or approval from the IT department (so-called “shadow AI”). At first glance, the focus is typically on efficiency. On closer inspection, however, potential patents, trade secrets and contractual obligations may also be affected.
Using AI Tools in Day-to-Day Business
Typical situations include inserting draft texts, presentations, source code or technical descriptions into a public AI tool in order to improve wording or identify errors. Employees often do not consider how the provider processes and stores this data.
Particularly in R&D, product design and sales, information that is economically highly sensitive for the company can be disclosed to external parties in this way.
Impact on Patents
For obtaining a patent, it is essential that the underlying invention has not been disclosed prior to filing—not even by the inventor. However, if technical solution approaches or even complete invention descriptions are fed into a public AI system, there is a risk that this may be considered a disclosure to a third party or may in fact become accessible to such a third party.
Trade and Business Secrets
Legal protection of trade secrets requires that appropriate confidentiality measures are in place. If internal information is disclosed to external AI services in an uncontrolled manner, this may, in the event of a dispute, be viewed as a failure to implement such protective measures.
This is particularly sensitive in the case of technical drawings, source code, algorithms, price lists, and customer or project data. Even if information is entered “just briefly for text improvement,” the protection of trade secrets may be called into question if a dispute arises.
Contractual Obligations towards Customers
Many projects are also subject to strict contractual confidentiality obligations. If employees process protected customer information using AI tools that have not been approved, this may constitute a breach of contract. In addition to potential legal liability, there is a risk of losing the trust of key business partners.
Companies should therefore clearly define whether, and to what extent, AI tools may be used in customer projects.
Practical Recommendations
Instead of imposing a strict prohibition, it is advisable to establish a clear framework for the use of AI:
- Define which information must not be entered into public AI systems (e.g. unpublished inventions, source code, confidential customer and project data).
- Provide secure AI solutions within the company, where appropriate backed by contractual safeguards.
- Offer short, practice-oriented training courses for employees on IP risks associated with the use of AI.
- Review and adapt invention disclosure forms, NDAs and project agreements to take the use of AI into account.
Conclusion
“Hidden” use of AI within a company is not a marginal issue; it directly affects industrial property rights. Anyone wishing to effectively protect patents, designs, trademarks and trade secrets should establish clear rules for the use of AI tools and embed them in day-to-day practice.
Canzler & Bergmeier supports companies in finding a sensible balance: using AI productively—without jeopardising their own IP rights.