When a rule becomes a target, it ceases to be a good rule. In the three years since the Supreme Court issued its opinion in Alice, there have been positive changes to patent applications, but there remains a long-term risk that patent practitioners will use tricks to beat the Alice test. Here, we focus on the changes to patent applications by drafters, as well as changes to patent applications that have issued since Alice.
In Alice v. CLS Bank, the Supreme Court laid down a straightforward, if difficult to consistently apply, test for patent eligibility under 35 U.S.C. § 101. Patent eligibility is to be determined based on 1) whether the claims at issue are directed to a patent-ineligible concept; and 2) if so, whether the claim’s elements, considered both individually and as an ordered combination, transform the nature of the claims into a patent-eligible invention. This test is intended to be used to determine if a claimed invention is abstract, and thus patent ineligible under § 101. We conclude that, in the short term, this test has resulted in positive changes to patent applications, as drafted. However, strategies by patent practitioners to circumvent this test may limit the test’s effectiveness in the future
Since Alice, there has been a sharp drop in allowance rates for patent applications classified into the business method technology center. Different strategies for patent preparation and prosecution in a post-Alice world have emerged. For example, commenters have proposed increasing the level of detail with which an invention is described in a patent application and including more detail in the claims. Our analysis focuses on patent applications that issued as patents from 2010-2017 to determine whether there have been any noticeable changes in patent language that may be attributed to post-Alice efforts to ensure eligibility under the Alice test.
In Figure 1, we looked at the average word count for the first independent claim in applications by issue year. Between 2010 and 2014, we found that the average word count in the first independent claim remained fairly consistent. For example, the difference between the highest average word count—in 2010, 257 words—and the lowest average word count—in 2013, 242 words—was a decrease of approximately 6%. Overall, the average word count in the first independent claim decreased from 257 words in 2010 to 247 words in 2014, representing a net decrease of approximately 4%. However, thereafter we saw increases of approximately 8.5% between 2015 and 2016 and 16% between 2016 and 2017 in the average word count for the first independent claim. Prior research has argued that an increase in the word count of the first independent claim corresponds to a narrowing of patent scope. Consequently, if the public policy goal undergirding the Alice decision was a “rebuke to… overbroad, poor-quality software patents,” we can conclude that the Alice decision has been successful, despite the aforementioned difficulties in its application.
Patent practitioners are not merely repeating the same words to increase the word count, they are adding significant quantities of unique words. In Figure 2, we looked at the average word count in the first independent claim of issued patents by unique word count. In other words, a repeated word would only count a single time in our analysis. We found that between 2010 and 2014, the average number of unique words in the first independent claim remained relatively stable between 73 and 71 words. Again, we saw an increase of approximately 19% in 2017 relative to the 2010 to 2014 baseline. Given the increase in the number of unique words, rather than merely repeating words in the first independent claim, we may conclude that patent practitioners are attempting to include additional detail in the claims, rather than simply repeating words to satisfy the pencil test.
The results for the average number of unique words in the first independent claim correlate to the results for the average word count in the first independent claim (described with respect to Figure 1). For the years from 2010 to 2014, the average number of unique words in the first independent claim decreased by approximately 3% (compared to approximately 4% for the average number of words in the first independent claim). In addition, for the years from 2015 to 2017 the average number of unique words in the first independent claim increased by approximately 15% (compared to a 19% increase for the years from 2015 to 2017 for the average number of words in the first independent claim).
We can also see patent practitioners attempting to include additional details in the specifications of patent applications, though the evidence for this is less clear. Previous analyses have reported that patent specification length has been increasing since long before 2010. As shown in Figure 3, from 2010 to 2013, the average patent application length was approximately 12,417 words. However, from 2015 and 2017, the average patent application length increased to over 14,700 words. Even though the average number of words in patent specifications has been increasing since 2010, the rates of increase are different for the years 2010 to 2014 (the year of Alice) and for the years 2015 to 2017 (the years following Alice). For the years 2010 to 2014, the rate of increase was approximately 242 words per year, or approximately 2% per year (with a decrease in average word count from 2012 to 2013). For the years 2015 to 2017, the rate of increase was approximately 504 words per year, or approximately 11% per year (with a decrease in average word count between 2014 and 2015). This may be a result of successful patents including increased numbers of benefit statements in patent specifications, thereby enabling patent practitioners to argue that the claims are directed to a technical solution to a technical problem. Alternatively, this may be a result of issued patents including additional disclosure details, additional fallback positions for prosecution, or additional implementation alternatives.
Another proposed strategy for ensuring eligibility under the Alice test is to avoid classification into technology center 3600 and/or class 705. A patent drafter may emphasize technical aspects of an invention in the first independent claim, abstract, and title to ensure classification into a favorable art unit. It has been noted that “substantially fewer recent filings [were] classified in the business method class (class 705) as compared to the time periods pre-dating the Alice decision.” An example of using this strategy may be to include words, such as “Artificial Intelligence” “Data Structures” “Digital Processing” and the like, in patent titles, abstracts, or first independent claims. This could increase a likelihood of classification into classes other than class 705. Such an increase, if the result of attempts to avoid classification into a more difficult art unit, undermines the utility of the Alice decision by increasing a likelihood that a patent drafter obfuscates the true purpose of an invention in the patent.
Preliminary data does show an increase in use of, for example, the term “artificial intelligence,” in patent titles, abstracts, and first independent claims; however, the sample size of applications filed post-Alice that include the term “artificial intelligence” is relatively small. Future investigation is warranted as more applications are published.
The aforementioned findings suggest both a positive outcome to establishment of the Alice test as well as a potential pitfall. Alice appears to have resulted in positive developments, such as narrowing the scope of claims and increasing disclosure of technical benefits of inventions in the specification. However, as patent practitioners become increasingly savvy with regard to avoiding eligibility challenges, there remains the potential for meaningless expansion to replace meaningful claim limitation, and for obfuscation to be used to avoid further public disclosure.
Author’s Note: This article would not have been possible without the assistance of Rocky Berndsen and Jaclyn Jones.