by Dennis Crouch
The recent eligibility decision in AI Visualize v. Nuance, __ F.4th __ (Fed. Cir. 2024), gives me pause to consider more general eligibility issues of AI Inventions. When does the design or creation of AI system elements qualify as an eligible invention? In his recent article, Prof. Nikola Datzov wrote what we have all been thinking: “Innovative applications of AI are everywhere we look [and are] revolutionizing our society.” Nikola L. Datzov, The Role of Patent (In)Eligibility in Promoting Artificial Intelligence Innovation, 92 UMKC L. REV. 1, 4 (2023).
In AI Visualize, the Federal Circuit sided with the accused infringer in finding the asserted claims ineligible under the two-step Alice framework. AI Visualize had asserted four related patents that facilitated use of a low-bandwidth web portal for visualizing 3D/4D medical scans. The key here is to use virtual views and a system to determine which views have already been downloaded. Some claims require a unique identifiable key for each view; others use a tiered approach – first sending lower-quality frames for immediate viewing followed by higher-quality frames. U.S. Patent Nos. 8,701,167; 9,106,609; 9,438,667; and 10,930,397. The district court dismissed the case void from the outset, finding asserted claims directed to patent-ineligible subject matter under 35 U.S.C. § 101. AI Visualize, Inc. v. Nuance Commc’ns, Inc., 610 F. Supp. 3d 638 (D. Del. 2022). On appeal, the Federal Circuit affirmed.
At step one, the court concluded the claims were directed to the abstract idea of “retrieving user-requested, remotely stored information.” Although the claims reciting the creation of “on the fly” virtual views, the court reasoned that “the claim language makes clear that virtual view ‘creation’ is achieved by the manipulation of a portion of the existing [data set].”
At step two, the Federal Circuit agreed with the district court that “the asserted claims involved nothing more than the abstract idea itself” and conventional computer technology. The court found that the creation of a virtual view was an abstract idea that was also known in the art, as conceded in the patent specifications and later at oral arguments. Therefore, that limitation could not supply the inventive concept required to transform the claims into patent-eligible subject matter. At oral arguments, AI Vis’s attorney Rajkumar Vinnakota provided further technical explanation, including an explanation of how “new frames are going to be created to fill in the gaps, essentially, for what the user requested. Those frames get transmitted back to the local processor, in the case of a table or something like that, and those are then combined with the stored frames.” (Oral arguments, 04:30-04:47). In its decision, however, the federal circuit concluded that these technical details were not part of the invention as claimed. Further, at oral arguments Chief Judge Moore explained her perspective that the specification “describ[es] the general idea of a user creating virtual views, but I don’t see any details about how the server actually implements them.” (Oral Args, 05:53-05:59).
The ultimate holding focused on how the claim language defined the “creation” of virtual views, has interesting implications for the eligibility of AI systems, particularly generative AI (GenAI) technologies. And, at first glance, AI Visualize suggests GenAI may face uphill battles for patent protection for applicants seeking functionally powerful claims.
A key factor will likely be how the “generation” of the new AI content is defined in the claims. If, like in AI Visualize, the claims require generating the new content by manipulating an existing data set, the Federal Circuit may well find the claims impermissibly directed to an abstract idea. The solution though is to recite a particular innovative technological solution for the creation of new content. The closer the claims come to reciting a concrete technical solution for how the AI improves computer functionality, the better the prospects for eligibility.
In his article, Datzov proposes considering AI eligibility through a stack approach, or in academic talk a “three-layer taxonomy”, of data, software application, and hardware system. As you might imagine, eligibility becomes as you delve deeper into the stack; although many computer hardware system claims have been deemed ineligible when they relied upon conventional computer components and where the true innovation was found in the functional data output. In other words, at all three levels, the key for eligibility is to expressly claim how the technology works, rather than just its function or objective. I have heard from patent attorneys that they are more than ever sending invention disclosures back to the inventors for more development until the technological improvement becomes apparent.
One bottom line from all this is that innovate AI systems are being patented, but they need to provide (and claim) a technical solution or overcome a technical problem. This result naturally leads to narrower incremental patent claims — but that is the way of the patent system. We don’t get 11 million patents that are each an archetypal constellation. Still, a second reality is that the current approach means that application layer developers are less likely to achieve patent eligible inventions, whereas those building the foundation models as well as those coordinating the linkage and layering of AI systems have a much greater chance. Finally, we have to recognize that AI continues to be a buzzword without tight definitions — by some counts more than 10% of US patent applications are AI related. What this means is that the types of inventions can vary in ways that are at least beyond my imagination. The result then is that creative and skillful patent drafting is more important than ever.
For all the patent huffers who post here relentlessly assuring us all that it’s the lack of patents (LOL!) that is slowing innovation in the “tech” sphere, here’s another view that hasn’t been warped by freebasing all that patent crack.
link to wheresyoured.at
Perhaps a summary…?
What someone needs to do is review the CAFC cases to see where they completely ignore the technical solution/technical problem concept.
Case in point: ADAPTIVE STREAMING INC. v. NETFLIX, INC. , No. 20-1310 (Fed. Cir. 2020)
From the actual patent (7,047,305):
Part of the problem: “Despite the availability of these modern day techniques, numerous limitations still exist. All of these techniques often have different formats, which make information difficult to transfer from one device type to another device type. Such formats are also difficult to control especially for a combination of audio and video information. Further, each of these techniques is often single purpose, which also limits use of each of these techniques. Additionally, video broadcasting for personal content information could not be distributed effectively, since many end user devices often used different formats and the like. Although there have been many advances, there are still numerous limitations, as noted.”
Part of the solution: “According to the present invention, techniques including a system for digital video processing are provided. In an exemplary embodiment, the present invention provides a system for personal broadcasting where the source audio/video and the displayed audio/video can be in a different format.”
In the court case, not a single mention of a problem that was addressed. Didn’t look at it.
Now, maybe that was the patent drafter’s issue: maybe they could have spent more time stating “For those CAFC judges, the technical problem is the following…. Various technical solutions to this include the following….”. But I’ve seen too many cases where this is ignored.
“ In an exemplary embodiment, the present invention provides a system for personal broadcasting where the source audio/video and the displayed audio/video can be in a different format.”
Wow what difficult problem to solve!!! What a baffling conundrum, almost impossible to conceive of the problem itself it’s just so outside the bounds of what we are used to thinking about.
I am so shocked this patent went up in flames.
“In other words, at all three levels, the key for eligibility is to expressly claim how the technology works, rather than just its function or objective.”
Can someone point me to an article that gives a half dozen concrete examples of claim language that works (and how you would modify it so it doesn’t work), or vice versa? Thank you.
No such thing – ‘Tis a Gordian Knot.
OT, but rarely does patent litigation occupy an international news magazine like The Economist. So I was surprised to see that the April 20th issue has a full page 23 article entitled “Niche Towns – If you build it, they will sue – Marshall and Waco – How two small Texas towns became the patent-law centre of America.” [Including some history, local business benefits, statistics, comments, and recent developments.] The included photo of a street om down-town Waco has a typical Economist humorous title: “Caution, lawyers crossing.”
The problem here seems to be typical claim overreaching of all claims, unrelated to AI, and falling into the Alice rabithole, as demonstrated by this part of the decision:
“AI Visualize points to multiple passages of the specification to support its view that “creation” of virtual views provides a technical solution to a technical problem, including one passage that addresses dynamic and static virtual views by describing how related image frames are selected from a VVD. .. As noted, we refuse to import details from the specification if those details are themselves not claimed. ChargePoint, 920 F.3d at 769. There is no recitation in the claim about how to create frames or virtual views, much less in a manner that would meaningfully support a technical solution to a technical problem in the prior art. We conclude that the asserted claims are directed to an abstract idea.”
“About HOW [my emphasis] to create”
That “how to” is what the specification is for – and need not (necessarily) be in the claim.
Right?
Wrong. Absolutely wrong and you should be ashamed.
We are not talking about a composition of matter claim, where the claim describes the structure of the composition and the method of making the composition is properly in the specification.
We are talking here about a method for transforming and transferring data on request. The “how” is EVERYTHING. On top of that, if the patentee is relying on the “how” to rescue a claim then those steps certainly must appear in that claim (or be incorporated into the claim by invocation of 112).
Hahahaha – go on and gaslight about how you are “pro patent.”
What if no one ever responded to anything it said here? Would it continue to exist?
This reminds me of the classic APJ question in patent application PTAB appeal oral arguments: “Mr. X, could you point out exactly where in the claims the distinctions you are arguing occur?” Somehow, too many attorneys never get the long repeated anchient lesson here that “we refuse to import details from the specification if those details are themselves not claimed” and keep making those arguments anyway. It may work with jurors, but not judges.
The real pity about this case, and many other cases that fall to §101, is that it’s clear from the document overall that the inventors solved a tough technology problem. The patent practitioner unfortunately failed to claim the technology solution of the invention, instead relying on claiming the output and abstracting away the inventor’s technology solution to the point that the claim is directed to an abstract concept.
“ it’s clear from the document overall that the inventors solved a tough technology problem”
Describe this “tough problem” and the allegedly inventive solution.
This is a common pitfall in patent drafting/prosecution. It is implicit for many patent practitioners that their objective must be to secure the broadest possible scope, because this is where they see their distinct added value, and claiming concrete features incl. field-of-use limitations is perceived as a defeat if not professional malpractice.
Who grabs too much loses all. Nothing new under the sun.
francis – the opposite would be (and in my view, should be) considered professional malpractice – under the adage of slippery slope and give and inch and they will take a mile, ANY relinquishing of less-than-proper scope will ONLY be met with the view that such encourages ever more attempts to winnow down allowed scope.
This is less about ‘grabs too much,’ and more about ‘eternal vigilance to maintain proper scope.’
“ ANY relinquishing of less-than-proper scope will ONLY be met with the view that such encourages ever more attempts to winnow down allowed scope.”
What an effin n u t c a s e.
Sure….
Try again to gaslight people how you are soooo NOT anti-patent.
The fact that you did not like my point proves my point.
Isn’t “first sending lower-quality frames for immediate viewing followed by higher-quality frames” over a low bandwidth channel something that was used in some NASA planet probes?
There is zero new “technology” in these claims. You can practically hear the patentee pounding the table and screaming “3D and 4D medical scans!” over and over as if that makes a difference.
Context MAY make a difference – cue your running away and not replying how your reformatting of your hard drive went with that Britney Spears CD.
Bottom line is that any “AI” focus here is swallowed up by the pre-existing quagmire and Gordian Knot of 101 juris(im)prudence, as opposed to my well-presented (actual) issues with AI and innovation factors.
Anon – I think I missed your presentation of AI and innovation factors. Can you restate those?
Please have one of your students gather them – they go back at least a decade.
“ The closer the claims come to reciting a concrete technical solution for how the data processing instructions improve computer functionality, the better the prospects for eligibility.”
Fixed for accuracy. As alluded to by Dennis, “AI” is a trendy buzzword for “computer”. There is nothing special about “AI” that will make a patent stronger. On the contrary, “AI” is conceptually old and all of its applications are obvious in the sense that, as with any logic process carried out by a computer, there is no kind of data (content-wise) that the computer can not “handle”. All claims involving “AI” raise the exact same red flags as all other information-processing claims.
In fact, if we believe what the proponents of “AI” say about “AI”, many claims involving “AI” will be even MORE suspect as the “learning-capable” computer will have written its own instructions which will need to be sussed out by the patent applicant. This task, presumably, will be less than straightforward. Barring that, the applicant will be forced to simply claim the result achieved and we all know how well that works.
Malcolm’s musings muted by his avoidance of (meaningfully) addressing the Grand Hall experiment.
Case in point: “use of a low-bandwidth web portal for visualizing 3D/4D medical scans. The key here is to use virtual views and a system to determine which views have already been downloaded. Some claims require a unique identifiable key for each view; others use a tiered approach”
Why is this deemed an example of “AI” when it sounds exactly like any other claim to a specifically programmed computer?
“to a specifically programmed computer?”