Artificial intelligence (AI) is playing an increasingly important role in technological developments. The mere use of AI does not, in and of itself, make a development patentable. What matters is whether and how the AI contributes to solving a technical problem.
What Information Is Required for Patent Applications for AI-Based Inventions
AI-based inventions may be patentable if they make a concrete technical contribution. Inventors should therefore first describe the technical problem that the AI solves. Examples include more accurate measurements, more robust detection, improved control, or reduced computing and memory requirements.
The technical functionality is also important. A description of the invention should therefore explain what data is collected, how it is processed, what role the respective AI model plays, and what output is generated. Flowcharts, system diagrams, or specific application examples are particularly helpful.
Information about the AI model and the training process can also be crucial. The specific training dataset does not always have to be fully disclosed. However, the application should make it clear what characteristics suitable training data should have and why the model achieves the desired technical effect.
Finally, the technical improvement should be substantiated as plausibly as possible. Suitable evidence includes, for example, test results, comparative data, measurement values, or a clear technical explanation. General statements regarding improved accuracy or performance resulting from the use of the AI model are generally insufficient.
Conclusion
Inventors should file for AI-based inventions early on, as soon as the technical problem, the data used, and the technical advantage are identifiable. The better these points are documented, the more effectively a robust patent application can be prepared.
We would be happy to work with you to determine whether your AI-based invention is patentable and what technical information still needs to be added for a successful patent application.
