Patents and Artificial Intelligence (AI)

The intersection of patents and artificial intelligence (AI) is a rapidly evolving area, raising fascinating legal, ethical, and technical questions. Here are some key points where they overlap:

1. Patenting AI Innovations
  • Algorithms & Models: Generally, mathematical algorithms, including AI algorithms, are considered abstract ideas and are not patentable per se. However, if they are tied to a specific, practical application (like improving medical diagnostics or autonomous driving), they may qualify for patents.
  • AI Hardware: Innovations in hardware designed to accelerate AI computations (like AI chips) are often patentable.
  • Applications of AI: Using AI in a novel way in specific fields (e.g., AI-driven drug discovery or AI-enhanced image recognition) can also be patentable.
2. AI as an Inventor
  • This is a hot topic. AI systems like DABUS (an AI system that created inventions without human intervention) have raised questions about whether AI can be recognized as an inventor.
  • Legal Stance: Most jurisdictions (like the US, EU, and UK) currently require a natural person to be named as the inventor. However, some countries, like South Africa and Australia, have been more open to recognizing AI-generated inventions in certain cases.
3. Challenges in Patent Law
  • Obviousness & Novelty: AI can generate solutions that are technically novel but might be seen as obvious because AI can process vast datasets and identify patterns faster than humans.
  • Disclosure Requirements: Patent applications must clearly explain how an invention works. With complex AI models (like deep learning networks), explaining the inner workings can be challenging due to their “black-box” nature.
4. AI in Patent Analytics

Are you looking for more on the legal side, technical applications, or something specific like how AI tools are reshaping patent law practices?

Soure- Swarupa Ghosh -Patent Attorney, Calcutta High CourtIPINDIA, Chatgpt