Farkas Szabolcs, president of the Hungarian Intellectual Property Office (SZTNH), warned on the Portfolio Business podcast that an increasing number of patent applications prepared with the help of artificial intelligence do not meet patentability requirements. He also cautioned that inventors risk exposing confidential information if they share details with public AI systems.
What is required to obtain a patent?
Farkas emphasized that a good idea alone is not enough for patent protection. A patent requires a technical solution that is genuinely new and practically implementable. Novelty is assessed globally, so what inventors disclose publicly before filing can affect the outcome.
How SZTNH uses AI
SZTNH already employs artificial intelligence for certain tasks, including searching for prior art and processing large volumes of documents. Farkas said the office expects to automate additional functions in the coming years.
Regulation, adaptation, and Hungary’s position
The discussion noted that technological development is currently outpacing regulation, but Farkas believes the patent system can adapt to the challenges posed by AI. He also pointed out that Hungary ranks 36th in the World Intellectual Property Organization’s (WIPO) innovation ranking covering 139 countries. Improving the country’s position, he argued, requires not only more inventions but also a more predictable economic and investment environment.
Why this matters for inventors
The podcast highlighted that AI brings both benefits and risks: it can speed up research and administrative work but can also lead to flawed or insufficient patent filings and accidental disclosure of confidential information. Inventors and companies should therefore use public AI tools cautiously and remain mindful of the technical and legal requirements of the patent system.
The Portfolio Business podcast episode is available on Spotify, Apple Podcasts and other major platforms.
