Patent applications for artificial intelligence (AI), machine learning, and other software-related inventions are often rejected by the US Patent and Trademark Office as being too abstract and not entitled to obtain a patent under 35 USC §101. The Patent Office recently issued a memo providing patent examiners (and furthermore patent applicants) to assess eligibility for software-related inventions. This memo contains important reminders that will help applicants seeking to patent software-related inventions avoid rejection.
As explained in detail on the Patent Office Subject Eligibility Web page, key research into patent eligibility involves two prongs. 1) whether the claim recites a judicial exception of eligibility (such as abstraction, mental processes, products of nature, etc.), and in that case, 2) whether the claim integrates judicial exceptions.
For Prong 1, the notes include the following helpful reminders:
For Prong 2, the notes include the following helpful reminders:
- Consider the claim Overall It claims the wording alone, not the wording of individual claims.
- Improved computer functionality or another technology or technical field is evidence of patent eligibility, as it is particularly important for AI, machine learning, and other software-related claims. The specification of a patent application does not require explicit discussion of improvements, but the specification can be consulted when determining whether the claimed invention provides improvements.
- Don't oversimplify your billing. This inquiry involves assessing whether the claim is recited or not. Ideas Solution (patent not eligible) or reciting Specific solutions Problem (patent qualification).
The memo also provides a general reminder that a claim should be denied under section 101 only if the claim is more than 50% likely to be ineligible for the patent. Of course, the criteria are abstract in itself and can provide the patent applicant's basis to claim that the claimed invention is patentable or not.
[View source.]
