The current patent system needs to be adjusted to keep up with the pace of development of artificial intelligence, to ensure a balance between encouraging innovation and preventing technological monopolies.
Changes could include improved review processes, evaluation standards for AI-related patents, and research exemption mechanisms to avoid stifling innovation.
- There is a need to update the standards and examination procedures for AI-related inventions.
- Balancing intellectual property protection and ensuring access and innovation.
- Practical solutions include fast-track, AI expert review, and research exemptions.
Why the patent system needs to evolve
The traditional patent system is often slow and difficult to adapt to the fast pace of innovation in the AI field.
Long review cycles, “novelty” and “breakthrough” standards that are difficult to apply to models and data; this can lead to overly broad or overlapping patents, stifling competition and research.
Key challenges with AI-related patents
Determining the subject of innovation, the scope of protection for algorithms and data, and assessing breakthroughs are complex issues.
Platform patents can create a “patent thicket” that creates legal barriers for startups and researchers, and slows down technology adoption and innovation.
Possible solutions to balance protection and innovation
Measures include updating due diligence standards, applying fast-track for AI applications, adding AI experts to the due diligence body, and XEM exemptions for research purposes.
Adding a license Chia mechanism, transparent standards for training data ownership, and clear guidance on the scope of protection helps reduce the risk of monopoly while still encouraging R&D investment.
How will the change affect startups?
Reforms could reduce regulatory barriers, making it easier for startups to access technology, but also require them to update their intellectual property policies and licensing strategies.
Can patent standards be modernized quickly?
Changing standards can be done incrementally — like fast-tracking, testing new standards — but it takes time to reach consensus and implement effectively.
What does a research exemption mean?
Research exemptions allow patents to be used for testing and development purposes without being sued, thereby promoting academic and applied innovation.
Where should stakeholders start?
Lawmakers, patent offices, and the research community should work together to develop specific guidelines for AI patents and test pilot mechanisms before scaling up.
