PatentNext Summary: AI-related inventions have experienced explosive growth. In view of this, the USPTO has provided guidance in the form of an example claim and an “informative” PTAB decision directed to AI-related claims that practitioners can use to aid in preparing robust patent claims on AI-related inventions.

The below article provides additional details.


Artificial Intelligence (AI) has experienced explosive growth across various industries. From Apple’s Face ID (face recognition), Amazon’s Alexa (voice recognition), to GM Cruise (autonomous vehicles), AI continues to shape the modern world. See Artificial Intelligence.

It comes as no surprise, therefore, that patents related to AI inventions have also experienced explosive growth.

Indeed, in the last quarter of 2020, the United States Patent and Trademark Office (USPTO) reported that patent filings for Artificial Intelligence (AI) related inventions more than doubled from 2002 to 2018. See Office of the Chief Economist, Inventing AI: Tracking The Diffusion Of Artificial Intelligence With Patents, IP DATA HIGHLIGHTS No. 5 (Oct. 2020).

During the same period, however, the U.S. Supreme Court’s decision in Alice Corp. v. CLS Bank International cast doubt on the patentability of software-related inventions, which AI sits squarely within.

Fortunately, since the Supreme Court’s Alice decision, the Federal Circuit clarified (on numerous occasions) that software-related patents are indeed patent-eligible. See Are Software Inventions Patentable?

More recently, in 2019, the United States Patent and Trademark Office (USPTO) provided its own guidance on the topic of patenting AI inventions. See 2019 Revised Patent Subject Matter Eligibility Guidance. Below we explore these examples.

USPTO Example 39 (“Method for Training a Neural Network for Facial Detection”)

As part of its 2019 Revised Patent Subject Matter Eligibility Guidance (the “2019 PEG”), the USPTO provided several example patent claims and respective analyses under the two-part Alice test. See Subject Matter Eligibility Examples: Abstract Ideas.

One of these examples (“Example 39”) demonstrated a patent-eligible artificial intelligence invention. In particular, Example 39 provides an example AI hypothetic invention labeled “Method for Training a Neural Network for Facial Detection” and describes an invention for addressing issues of older facial recognition methods that suffered from the inability to robustly detect human faces in images where there are shifts, distortions, and variations in scale in scale and rotation of the face pattern in the image.

The example inventive method recites claim elements for training a neural network across two stages of training set data so as to minimize false positives for facial detection. The claims are reproduced below:

     1.     A computer-implemented method of training a neural network for facial detection comprising: 

     collecting a set of digital facial images from a database; 

     applying one or more transformations to each digital facial image  including mirroring, rotating, smoothing, or contrast reduction to create a modified set of digital facial images; 

     creating a first training set comprising the collected set of digital facial images, the modified set of digital facial images, and a set of digital non-facial images; 

     training the neural network in a first stage using the first training set

     creating a second training set for a second stage of training comprising the first training set and digital non-facial images that are incorrectly detected as facial images after the first stage of training; and 

     training the neural network in a second stage using the second training set.

The USPTO’s analysis of Example 39 informs that the above claim is patent-eligible (and not “directed to” an abstract idea) because the AI-specific claim elements do not recite a mere “abstract idea.” See How to Patent Software Inventions: Show an “Improvement”. In particular, while some of the claim elements may be based on mathematical concepts, such concepts are not recited in the claim. Further, the claim does not recite a mental process because the steps are not practically performed in the human mind. Finally, the claim does not recite any method of organizing human activity, such as a fundamental economic concept or meaning interactions between people. Because the claims do not fall into one of these three categories, then, according to the USPTO, then the claim is patent-eligible.

PTAB designates as “Informative” a Decision regarding an AI invention

As a further example, the Patent Trial and Appeal Board (PTAB) more recently applied the 2019 PEG (as revised) in an ex parte appeal involving an artificial intelligence invention. See ex parte Hannun (formerly Ex parte Linden), 2018-003323 (April 1, 2019) (designated by the PTAB as an “Informative” decision).

In Hannun, the patent-at-issue related to “systems and methods for improving the transcription of speech into text.” The claims included several AI-related elements, including “a set of training samples used to train a trained neural network model” as used to interpret a string of characters for speech translation. Claim 11 of the patent-at-issue is illustrative and is reproduced below:

      11.     A computer-implemented method for transcribing speech comprising: 

     receiving an input audio from a user; normalizing the input audio to make a total power of the input audio consistent with a set of training samples used to train a trained neural network model; 

     generating a jitter set of audio files from the normalized input audio by translating the normalized input audio by one or more time values; 

     for each audio file from the jitter set of audio files, which includes the normalized input audio:

     generating a set of spectrogram frames for each audio file; inputting the audio file along with a context of spectrogram frames into a trained neural network; obtaining predicted character probabilities outputs from the trained neural network; and 

     decoding a transcription of the input audio using the predicted character probabilities outputs from the trained neural network constrained by a language model that interprets a string of characters from the predicted character probabilities outputs as a word or words. 

Applying the two-part Alice test, the Examiner had rejected the claims finding them patent-ineligible as merely abstract ideas (i.e., mathematical concepts and certain methods of organizing human activity without significantly more.)

The PTAB disagreed. While the PTAB generally agreed that the patent specification included mathematical formulas, such mathematical formulas were “not recited in the claims.” (original emphasis).

Nor did the claims recite “organizing human activity,” at least because, according to the PTAB, the claims were directed to a specific implementation comprising technical elements including AI and computer speech recognition.

Finally, and importantly, the PTAB noted the importance of the specification describing how the claimed invention provides an improvement to the technical field of speech recognition, with the PTAB specifically noting that “the Specification describes that using DeepSpeech learning, i.e., a trained neural network, along with a language model ‘achieves higher performance than traditional methods on hard speech recognition tasks while also being much simpler.’”

For each of these reasons, the PTAB found the claims of the patent-at-issue in Hannun to be patent-eligible.


Each of Example 39 and the PTAB’s informative decision of Hannun demonstrates the importance of properly drafting AI-related claims (and, in general, software-related claims) to follow a three-part pattern of describing an improvement to the underlying computing invention, describe how the improvement overcomes problems experienced in the prior art, and recite the improvement in the claims. For more information, see How to Patent Software Inventions: Show an “Improvement”

Subscribe to get updates to this post or to receive future posts from PatentNext. Start a discussion or reach out to the author, Ryan Phelan, at or 312-474-6607. Connect with or follow Ryan on LinkedIn.