Description

Between the years 2014 to 2023, a total of 54,000 Generative Artificial Intelligence related patent families were filed, and the top five inventors are located in China (38,210), the United States (6,276), the Republic of Korea (4,155), Japan (3,409), and India (1,350).

Image and video data dominate GenAI patents (17,996 inventions), followed by text (13,494 inventions) and speech/music (13,480 inventions), while molecules rapid growth is experienced in molecule, gene and protein-based data (1,494 inventions since 2014), showing a 78% average annual growth rate over the past five years.

Now, despite this unprecedented surge in GenAI patent filings, the fundamental legal infrastructure governing AI inventorship, disclosure requirements, and subject matter eligibility remains fundamentally misaligned with the technological reality these patents represent.

This requires one to ask some questions on the global AI patentability landscape per se.

• How can patent offices meaningfully examine 54,000 GenAI patent families when basic questions of AI inventorship remain unresolved, particularly given that current patent regimes recognize only human inventors while AI systems increasingly generate novel solutions independently?
• Why are innovators filing thousands of patents in jurisdictions like China and the US when trade secret protection may offer superior protection for AI innovations that remain commercially concealed and difficult to reverse-engineer?

• What does it mean that image/video data dominates GenAI patents (17,996 inventions) when copyright law fundamentally challenges AI’s use of training data, creating a paradox where patent protection may be undermined by copyright infringement in the underlying training process?
• How sustainable is this patent filing surge when AI-generated content often lacks copyright protection due to human authorship requirements, potentially rendering the commercial value of these patents questionable?
• Given that molecules/genes/proteins show 78% annual growth in AI patents, how do patent offices assess the inventive step and non-obviousness when AI can explore vast molecular spaces that would take humans years to investigate?

While it may not be possible to have direct answers to such questions – this handbook, developed for the first time in India, addresses the understanding around Patent Law in the context of Artificial Intelligence technologies, beyond the usual public wisdom of the generative AI space, coupled with the market hype around it.

The handbook therefore explains:

• The AI and Copyright Law Intersection
• • Whether facets of the Fair Use Doctrine in AI Training have any effect on AI Patentability issues

• Types of AI-related Patents
• Core Technical and Legal Issues including those of Inventorship Determination, Non-obviousness, Classification Consistency issues, and others
• Ownership and Knowledge Management issues
• Data Governance Contexts in AI Patentability

Disclaimer

This handbook uniquely combines mind maps and infographics to explain complex AI and intellectual property concepts.

Unlike extensive encyclopedias or policy trackers, it’s not simply a repository of information; rather, it aims to provide a more complete and nuanced understanding of the legal challenges surrounding the use, involvement, and democratisation of AI in various classes of intellectual property, addressing potential gaps or biases in other sources.

The Global AI Inventorship Handbook’s relevance and impact aren’t limited by the hype surrounding Generative AI and language models, ensuring its value remains unaffected by fleeting trends.