USPTO’s 2024 Guidance Rules Out Patents for Data Harvesting Methods
USPTO’s 2024 Patent Subject Matter Eligibility Guidance excludes data harvesting methods from patent eligibility.
Why it matters: Legal professionals must adjust patent and litigation approaches for AI innovations, as data harvesting methods are now explicitly non-patentable. This affects drafting strategies and intellectual property protections amid evolving patent law.
- The USPTO's 2024 Patent Subject Matter Eligibility Guidance took effect July 17, 2024, clarifying AI and data-related patent criteria.
- Data harvesting methods, viewed as abstract processes, are explicitly excluded from patent eligibility per the new guidelines.
- The U.S. Supreme Court’s key rulings (Alice Corp. v. CLS Bank, 2014; Mayo v. Prometheus, 2012) underpin this exclusion, limiting patents on abstract ideas and natural phenomena.
- 35 U.S.C. §101 defines patent-eligible subject matter as useful processes, machines, manufactures, or compositions of matter, excluding abstract ideas.
- Legal commentary notes that patent claims on AI training data collection must rely on innovations beyond mere data gathering processes.
The U.S. Supreme Court has long established that abstract ideas, laws of nature, and natural phenomena are not patentable under 35 U.S.C. §101. Notable cases including Alice Corp. v. CLS Bank (2014) and Mayo Collaborative Services v. Prometheus Labs (2012) have clarified this framework to prevent patents that claim abstract concepts without inventive application.
On July 17, 2024, the USPTO’s Patent Subject Matter Eligibility Guidance Update formally incorporated these principles, explicitly excluding data harvesting and AI training data gathering methods from patent eligibility. The guidance explains that data harvesting is an abstract data collection process, which alone does not meet the requirement of a "useful process" or a new machine or manufacture.
This means patent applications focusing solely on methods of collecting or using data without additional inventive steps will likely be rejected. Such restrictions reflect a consistent judicial and administrative pattern aiming to keep basic data-related processes in the public domain.
Legal experts emphasize the importance of emphasizing concrete technological improvements or novel applications in AI-related patents rather than claiming broad data processing or harvesting methods. Intellectual property strategies should prioritize protecting algorithms, machine implementations, and specific technological innovations beyond mere data collection.
Corporate counsel and patent attorneys should reassess claims related to AI training datasets and data-handling processes to avoid subject matter eligibility rejections. This update directly impacts litigation risk assessments and patent portfolio management for companies innovating in AI and big data analytics.
By the numbers:
- July 17, 2024 — Effective date of USPTO’s Patent Subject Matter Eligibility Guidance update
- 2014 — Year of Supreme Court’s Alice decision limiting patentability of abstract ideas
- 35 U.S.C. §101 — U.S. law defining patent-eligible subject matter
Yes, but: While data harvesting methods alone are excluded, patents may still cover specific technical implementations or novel AI models that incorporate such data in inventive ways.
What's next: Legal professionals should monitor how USPTO examiners apply this guidance in patent prosecution and await any further clarifications from federal courts on AI-related patent eligibility.