top of page
IP Tech & Innovation Services Annual (6).png

THE GEN AI RACE: AUTOMATION OF PATENTS

Updated: Jul 15

The Global GenAI Patent Hoarding Race

The volume of patent filings associated with generative AI has experienced unprecedented growth in recent times. According to data from the World Intellectual Property Organisation (WIPO), newer GenAI patents were published in 2024 and 2025 combined than in the entire preceding decade. Published GenAI patent families rose from roughly 14,000 in 2023 to over 37,800 by 2025, now accounting for nearly 9% of all AI-related patent publications globally. [1]

Key Corporate Actors

The race to secure these foundational technologies is no longer restricted to traditional software developers, expanding heavily into telecommunications, finance, and digital infrastructure:

  • SoftBank (Japan): Emerging as the global leader, the multinational conglomerate published nearly 3,000 GenAI patent families during the 2024–2025 period.

  • Chinese Conglomerates: Firms such as Tencent Holdings, Ping An Insurance Group, and Baidu closely follow SoftBank, driving a massive wave of domestic filings.

  • United States Tech Giants: Alphabet, Microsoft, and IBM remain highly competitive, anchoring the Western market's patent portfolio.

Geographical Distribution[2]

  • China: Remains the largest source of GenAI patent publications by a substantial margin, accounting for over 43,000 patent families in 2024 and 2025, with a compound annual growth rate (CAGR) of 64%.

  • United States: Recorded a 92% CAGR over the same period as commercial developments translated rapidly into formal filings.

  • Japan: Registered the highest relative growth among major inventor hubs, with a 210% CAGR.

2. AI Automating the Patent Legal Sector[3]

Within law firms and corporate legal departments, generative AI has transitioned from a novel, administrative drafting aid into the operational core of patent practice. Specialised legal-tech platforms, such as Patlytics, DeepIP, Solve Intelligence, and IP Author, are automating highly complex, time-intensive tasks historically reserved for junior associates and patent agents.

AI is predominantly automating three key phases of the patent lifecycle:

Prior Art Searches

Determining whether an invention is truly novel traditionally required patent attorneys to spend hours or days manually parsing global patent databases.

  • Specialised AI models can now ingest a technical disclosure and scan millions of documents across global patent and non-patent literature in minutes.

  • These systems automatically generate semantic similarity scores, flag potential conflicts, and compile novelty summaries with high accuracy.

Automated Patent Drafting

Writing a patent application requires precise legal terminology and rigid formatting.

  • Legal-tech platforms now simulate inventor interviews to capture technical parameters, subsequently generating complete first-draft patent applications, including claims, detailed descriptions, and abstract text, in under 20 minutes.

Certain tools can also auto-generate supporting figures, such as workflow diagrams and flowcharts, directly from a text description.

  


Office Action Responses

When a patent examiner rejects or objects to a patent application, lawyers must draft a formal response.

  • AI systems are now deployed to read the examiner's objections, analyse the prosecution history of the application, and draft targeted arguments and claim amendments to overcome the rejection.

  • This automation reduces the time required to draft an office action response by an estimated 70% to 80%.[4]

The Paradigm of Humans in the Loop

Despite the speed of these automated platforms, the legal industry has not fully displaced human practitioners. Because patent applications carry immense long-term financial consequences and must navigate strict compliance standards, such as data governance and client confidentiality, firms primarily utilise AI in a "hybrid" or "human-in-the-loop" model. Attorneys act as critical editors, validating the AI's technical output and legal arguments to ensure accuracy before submitting documents to patent offices.

 

References :

[1] (July 14, 2026). GenAI Innovation Soaring, With Patent Activity Nearly Tripling in Two Years. World Intellectual Property Organization. https://www.wipo.int/pressroom/en/articles/2026/article_0012.html

[2] (July 14, 2026). GenAI Innovation Soaring, With Patent Activity Nearly Tripling in Two Years. World Intellectual Property Organization. https://www.wipo.int/pressroom/en/articles/2026/article_0012.html

[3] (April 8, 2026). Patlytics Raises $40 Million Series B to Expand the AI Platform Purpose-Built for IP Work. Business Wire. https://www.businesswire.com/news/home/20260408770722/en/Patlytics-Raises-%2440-Million-Series-B-to-Expand-the-AIPlatform-Purpose-Built-for-IP-Work [4] Hahn, R. (2026). How Patent Attorneys Save 2.5 Hours Per Office Action Response with AI. ABIGAIL. https://abigail.app/blog/guides/patent-oa-response-time-savings


Comments

bottom of page