On Alice Rejections per USPTO Technical Center

By Michael Borella --

The U.S. Patent and Trademark Office (USPTO) organizes its examining corps into technical centers (TCs).  Each TC is dedicated to one or more general technical fields.  In some cases, one TC may include two or more unrelated fields, while some fields (software, for example) are represented in multiple TCs.  Currently there are nine TCs, not counting one dedicated to reissues.

Each TC is further subdivided into art units (usually 4-10) and examiners are assigned to these art units.  Thus, examiners tend to focus their efforts on examining patent applications within a fairly narrow field.

Over the nine years since 2014's Alice Corp. v. CLS Bank Int'l Supreme Court case, the number of 35 U.S.C. § 101 rejections has dramatically increased throughout the USPTO.  Anecdotally, however, many practitioners believe that certain TCs are much more likely to result in 101 rejections than others.  In practice, it is generally accepted that TC 3600 -- the TC to where most business-related inventions are routed -- issues the most Alice-based 101 rejections.

We tested this hypothesis by gathering data on Alice rejections per TC from 2020.  We had to look back this far due to the USPTO's 18-month publication delay causing not all relevant data from 2021 to be available.  Here, an Alice rejection is a rejection of at least one claim on the grounds of the claim(s) not meeting the judicial patent eligibility requirements of Alice.  This is a slightly narrower scope than § 101 rejections in general, which include claims rejected on grounds of failing to meet the statutory requirements and grounds of double patenting.

While the raw numbers of Alice rejections per TC in 2020 were not hard to obtain, they were also not very insightful when considered in isolation.  To be able to compare Alice rates across TCs, we also collected a count of § 103 rejections per TC in 2020.  By considering the ratio of § 103 rejections to Alice rejections, we are able to obtain a clearer understanding of how "difficult" each TC is with respect to § 101.  Particularly, using § 103 rejections (which are quite common) as a baseline, we can get a clear understanding of how many Alice rejections to expect per § 103 rejection.

As shown above, the ratio of § 103 to Alice rejections varies wildly across the TCs.  For instance, in TC 1600, you can expect one Alice rejection for about every fifteen § 103 rejections.  On the other hand, in TC 3600, you can expect one Alice rejection for about every three § 103 rejections.  Indeed, TC 3600 has the highest likelihood of an Alice rejection by far when compared to § 103 rejection volume.  This appears to confirm our anecdotal observations.

Other TCs that have relatively high rates of Alice rejections are 1600 and 2100.  Both of these TCs receive a high volume of software applications.  In contrast, TCs 1700, 2600, and 2800 have the lowest rates, with 1700 being quite low.

Based on experience, we expected a higher rate of Alice rejections in TC 3700.  In particular, gaming inventions get routed to this TC, and we have seen frequent Alice rejections for their applications.  This TC, however, also examines mechanical and medical inventions, which typically are not subject to § 101 challenges as often.  Thus, across TC 3700 as a whole, Alice rates fall about in the middle of the field.  This suggests that it might be informative to further break this data down by art unit in a follow-on exercise.

A word of caution -- while this data suggests that TC is highly influential on whether an application is going to receive a § 101 rejection, there are other factors.  By their very nature, some inventions (e.g., business methods, data processing) are more prone to a § 101 rejection than others (e.g., mechanical devices, chemical compounds).  Inventions with a high likelihood of triggering an Alice analysis may receive a § 101 rejection regardless of its TC.

Still, if a goal is to avoid TCs with higher Alice rates, this data provides a compelling view of just how much each TC relies on Alice rejections when compared to § 103 rejections.  As such, TC prediction and steering tools can be advantageous when used in conjunction with the numbers above.

Comments

[Astra]

Business Methods should be 31.3 not 3.13.

Interesting data otherwise.

[Michael Borella]

Astra,

74663 / 23882 = 3.126329

Mike

[Fact Checker]

"Each TC is further subdivided into art units (usually 4-10) and examiners are assigned to these art units." ??? Oops, TCs are divided into Art GROUPS each run by one or more TC Directors, which are then further divided into Art Units each run by one or more Supervisory Patent Examiners. There are many more than the stated 10 Art Units in the typical TC. For example, TC 1600 has 7 main groups and a total of 48 main units at about 8-30 examiners per unit: https://ptoweb.uspto.gov/patents/1600/

[Fact Checker]

If 103 rejections were intended as the reference or basis for normalization, then why would you report 103 in the numerator and 101 in the denominator? To avoid fractions? Seriously? Better to multiple by one hundred or otherwise create a score than to write about a relationship inverse to the intent of the article. This creates the awkward result of writing "TC 3600 has the HIGHEST likelihood of an Alice rejection by far when compared to § 103 rejection volume," but the LOWEST table value of "3.13."

The entire value of this article, such as it is, is captured in the implicit sentence, "...TCs that have relatively high rates of Alice rejections... receive a high volume of software applications." Hardware, wet steps, machines, manufactures, etc. tend to implicate Alice less often, while data processes do so more often. It's that simple. Of course, this doesn't mean that software can't be patented (e.g. Enfish, McRO, etc.), but it does mean that the courts have created an additional filter regarding software patents, among other types of judicial exception. Looked at another way, if there is a possible judicial exception, then to what extent does the technology lend itself to integrating that exception into the real, physical world as a practical application? Who should receive the benefits of a patent, and what should that patent contribute to the actual technology field and the economy?

To write that "TC" is an interesting variable correlating with the rate of Alice rejections, misses the obviously more relevant variables of technology type and even simply the balance of calculational process versus hardware, machine, manufacture, etc. regardless of technology type. Beginning an Alice inquiry with a focus on "TC" is like beginning an average annual temperature inquiry by asking about which U.S. state, in which clearly there's a more relevant variable, i.e. latitude. Why not just start with latitude? All of this is alluded to in passing near the end of the article, but it would seem to have made more sense, before posting the article, to have simply revised according to that penultimate paragraph. Instead, the closing sentence appears misleading. Despite TC predictions and "steering tools," it's the underlying technology that properly triggers Alice analysis, not attempts to game the assigned TC.

Also, what about outcomes? Alice is not a bar to patentability. It is a hurdle, just like each other statute. Which technologies receive the most applications? How does the final rate of allowable claims vary by technology? TC data may have been available, but it was a poor choice for this analysis.

[Michael Borella]

FC:

Not sure if you practice in this field, but when you say "Despite TC predictions and 'steering tools', it's the underlying technology that properly triggers Alice analysis, not attempts to game the assigned TC" I have to disagree. This is not at all what we see. TC, AU and examiner proclivities are highly determinative of whether you receive a 101 rejection and how hard it is to overcome.

I've spoken with examiners who feel like their AU and TC heads instruct them to issue and maintain 101 rejections. Other examiners have told me that the 101 decision is entirely up to them.

In an ideal world, the tech would be the determining factor in the eligibility analysis. But this is the real world and the real world is messy. The USPTO very much so.

And this analysis was performed with data that is publicly available and that readily lends itself to analysis. If you have a better methodology, feel free to carry it out and post the results in a comment. I'd be interested in seeing that.

[Fact Checker]

Mike,
Good points above. Re when to write and maintain Alice rejections, I've also had the experience of various policies being set at the PTO, varying over time and yes by group, unit and examiner. Sometimes such policy is unwritten and more of a tone. At least in TC1600, most often this seems either to come at the TC level, which often appears to be driven by OPLA (the internal patent law team), or to vary at the individual examiner level.
Thanks for your analysis. At the very least it stimulates discussion, using available data as you mention.
FC

[skeptical]

Mr. Borella,

Enjoyed your presentation today.

[Michael Borella]

Thanks Skeptical.