DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Applicant's arguments with respect to 35 U.S.C. 103 in regards to claims 1, 9 and 17 have been considered, however are not found to be persuasive due to the following reasons. Examiner respectfully disagrees with Applicant’s arguments because Gray teaches receiving a user query through a UI, generating LLM input/prompting to answer the query, producing an NL summary response, and using context from the client device, including “current or recent interaction(s),” “a current state of a query session,” the active foreground application, the current or recent state of that application, and content currently or recently rendered by the application. This is reasonably interpreted as a context of a processing state of the computing system, including prior processing that contributed to the UI state when the query was entered. Gray also teaches suppressing rendering of an LLM-generated summary when confidence/objective conditions indicate the summary should not be provided, and rendering it when the conditions are satisfied. Thus, the amended “context,” “valid/invalid,” and “display or omit” features are not absent from Gray as the Applicant argues.
Applicant’s argument is also not persuasive because the rejection relies on Inan for the second LLM evaluation feature, not on Gray alone. Inan teaches Llama Guard, an LLM based safeguard model for human AI conversations, which classifies LLM-generated responses, receives conversation data including a user message and an agent response, and outputs a binary “safe” or “unsafe” decision. That is an evaluation of the first LLM response and provides the claimed validity verdict. Gray further teaches that multiple LLMs may be truly different, including LLMs with different training, fine-tuning or different architectures, and Inan’s Llama Guard is a Llama2-7B model instruction-tuned for response classification. Therefore, it would have been obvious to use Inan’s safeguard LLM in Gray’s system to evaluate the first LLM’s response before the UI displays or suppresses it. The Applicant’s remarks mainly argue that no signal reference uses the exact claim language but that does not overcome the obviousness rejection based on the combined teachings of Gray and Inan as set forth below.
Applicant's arguments with respect to 35 U.S.C. 101 Abstract Idea in regards to claims 1-20 have been considered, however are not found to be persuasive due to the following reasons. Examiner respectfully disagrees with Applicant’s arguments because the claims still focuses on an information-evaluation process: receive a query, generate prompts, get an LLM answer, evaluate that answer with another LLM using context, decide whether the answer is valid, and display or omit it. That is still collecting information, analyzing information, making a judgement, and presenting a result. The added UI, processor, first LLM, second LLM, and “context of a processing state” do not by themselves show a practical application because the claim does not recite a specific improvement to the UI, processor, LLM architecture, model training, memory usage or other computer technology. Under USPTO guidelines, a claim that recites an abstract idea must include additional elements that integrate the idea into a practical application or add significantly more.
Applicant’s reliance on a “second LLM having different algorithmic structure” is also not enough. The claim only states the result, using a different LLM to evaluate the first LLM response, but does not claim how that different structure technically improves the computer system or the LLM itself. The Dec. 5, 2025 USPTO update and Ex parte Desjardins support eligibility where the claim actually reflects a technological improvement, including to machine learning or computer functionality, but they do not make all AI/LLM claims patent eligible. Here, the alleged improvement is mainly better answer screening or better informational accuracy, which is the result of the abstract evaluation process, no a claimed technical solution. Therefore, the arguments do not overcome the 101 abstract idea rejection.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101.
Claims 1, 9 and 17 are directed to an abstract idea. The claim core invention is essentially information processing and evaluation: generating prompts, obtaining textual responses, determining context, and then making a “validity verdict” and selectively presenting/withholding that response. Mental processes include concepts performed in the human mind such as observation, evaluation, judgment, opinion, and furthermore, the claims recite generic computer components.
The claims, as written does not clearly integrate that abstract idea into a practical application because the additional elements are stated at a high level of generality (processor(s), UI, first LLM and second LLM, generic prompts, and a generic computing system state). The claim does not recite a particular machine implementation or a specific improvement to computer functionality.
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the claims are (i) mere instructions to implement the idea on a computer, and/or (ii) recitation of generic computer structure that serves to perform generic computer functions that are well-understood, routine, and conventional activities previously known to the pertinent industry. Viewed as a whole, these additional claim element(s) do not provide meaningful limitation(s) to transform the abstract idea into a patent eligible application of the abstract idea such that the claim(s) amounts to significantly more than the abstract idea itself. Therefore, the claim(s) are rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter. There is further no improvement to the computing device.
Dependent claims 2-8, 10-16 and 18-20 further recite an abstract idea performable by a human and do not amount to significantly more than the abstract idea as they do not provide steps other than what is conventionally known in natural language processing.
Claims 2, 10 and 18: a mental process/information organization step implemented on a generic computer.
Claims 3 and 11: an abstract evaluation/judgment concept performed using generic computing components.
Claims 4 and 12: a mental process implemented on generic hardware.
Claims 5 and 13: does not integrate the abstract idea into a practical application.
Claims 6, 14 and 19: generic computer of the abstract idea without improving computer functionality.
Claims 7, 15 and 20: an abstract evaluation/judgment step performed by a generic computer.
Claims 8 and 16: the computing system to a tax calculation engine merely applies the abstract idea to a particular field (tax processing) without adding a technological improvement or inventive concept.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-7, 9-15 and 17-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Gray et al. (US 2024/0220735) in view of Inan et al. (“Llama Guard: LLM based Input-Output Safeguard for Human-Ai Conversations”; Dec. 7, 2023).
Claims 1, 9 and 17,
Gray teaches a method comprising: receiving, by at least one processor, a user query entered through a user interface (UI) (Gary teaches a receiving a query “based on user interface input” and gives typed, voice, image, and multimodal query examples [0051] [Fig. 2]; Gray also teaches a query “formulated based on user input” detected by a user input engine, including typed input, touch-screen selection, spoken input, and image input [0034]);
generating, by the at least one processor, a first prompt including at least the user query and configured to elicit an answer to the user query (Gray gives a prompt example: “answer [query]” [0014]; Gray also teaches a prompt in the form “In the context of <query>, summarize…” processed by the LLM [0007]);
inputting, by the at least one processor, the first prompt to a first large language model (LLM) having a first algorithmic structure and receiving a first response from the first LLM (Gary teaches that an LLM may be “sequence-to-sequence,” “transformer-based,” and may include “an encoder and/or a decoder,” with examples including PaLM and LaMDA [0016]; Gray also teaches processing the prompt/input using the LLM to generate LLM output reflecting an NL summary [0065-0067] [Fig. 2]);
determining, by the at least one processor, a context of a processing state of a computing system comprising prior processing contributing to a state of the UI at a time the user query was entered (Gray teaches a context engine that determines context from “current or recent interaction(s),” “a current state of a query session,” recent queries, profile data, location, active foreground application, the “current or recent states” of the active application, and content currently/recently rendered [0036]);
a second algorithmic structure different from the first algorithmic structure (Gary teaches that a first LLM may have “a first architecture” that differences from “a second architecture” of a second LLM [0139]).
generating, by the at least one processor, the answer to the user query and causing display of the answer by the UI, wherein the answer includes the first response for a valid verdict or omits the first response for an invalid verdict (Gray teaches rendering the NL based summary when confidence thresholds are met, but when thresholds fail, “redarning.. can be suppressed completely” [0072]; Gray also teaches causing the NL based summary to be rendered graphically in the UI [0073]).
The difference between the prior art and the claimed invention is that Gray does not explicitly teach generating, by the at least one processor, a second prompt including at least the context and the first response; inputting, by the at least one processor, the second prompt to a second LLM and receiving a second response from the second LLM comprising an evaluation of the first response; determining, by the at least one processor, a validity verdict of the first response using at least the evaluation of the first response from the second response.
Inan teaches generating, by the at least one processor, a second prompt including at least the context and the first response (Inan teaches that the evaluator task contains “a conversation” that can include “a single user message followed by a single agent response” [3.1] [pg. 3]; Inan’s Fig. 1 shows a response-classification prompt that includes a user message and an agent response, then asks for a safety assessment of the agent response [Fig. 1] [pg. 4]);
inputting, by the at least one processor, the second prompt to a second LLM and receiving a second response from the second LLM comprising an evaluation of the first response (Inan teaches “Llama Guard, an LLM-based input-output safeguard model” for classifying prompts and responses [Abstract] [pg. 1]; Inan further teaches that Llama Guard classifies “responses generated by LLMs,” which it calls “response classification” [Abstract] [pg. 1]);
determining, by the at least one processor, a validity verdict of the first response using at least the evaluation of the first response from the second response (Inan teaches that Llama Guard outputs “safe” or “unsafe” and that this output supports a “binary decision” [3.1] [pg. 3-4] [Fig. 1]).
Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the teachings of Gray with teachings of Inan by modifying the generative summaries for search results as taught by Gray to include generating, by the at least one processor, a second prompt including at least the context and the first response; inputting, by the at least one processor, the second prompt to a second LLM and receiving a second response from the second LLM comprising an evaluation of the first response; determining, by the at least one processor, a validity verdict of the first response using at least the evaluation of the first response from the second response as taught by Inan for the benefit of encouraging researchers to further develop and adapt them to meet the evolving needs of the community for AI safety (Inan [Abstract]).
Claims 2, 10 and 18,
Gray further teaches the method of claim 1, wherein the first prompt further includes at least one instruction for responding to the user query, the context, or a combination thereof ([0007] [0012-0014] a prompt of answer [query] can be processed using the LLM in generating the NL based summary; a prompt of “In the context of <query>, summarize <Content A>, <Content B>, <Content C>, and <Content D>” can be processed using the LLM to generate the NL based summary).
Claims 3 and 11,
Inan further teaches the method of claim 1, wherein the second prompt further includes at least one evaluation step, at least one inaccuracy criterion, or a combination thereof ([3.1] each task takes a set of guidelines as input, which consist of numbered categories of violation, as well as plain text descriptions as to what is safe and unsafe within that category).
Claims 4 and 12,
Gray further teaches the method of claim 1, wherein determining the context comprises: determining the processing state of the computing system ([0036] current state of a query session);
determining at least one data entry applicable to the processing state ([0036] profile data, and/or a current location); and
defining the context as data describing at least a portion of the processing state and the at least one data entry ([0036] determine a current context based on a current state of a query session (e.g., considering one or more recent queries of the query session), profile data, and/or a current location of the client device 110).
Claims 5 and 13,
Gray further teaches the method of claim 1, further comprising: determining, by the at least one processor, the processing state of the computing system by obtaining data from the computing system ([0036] a current state of a query session (location/profile data of the client device 110 (a processor and memory)));
wherein the computing system is separate from, and in communication with, at least one device comprising the at least one processor ([Fig. 1] [0038] one or more of the software applications can be hosted remotely (e.g., by one or more servers) and can be accessible by the client device 110 over one or more of the networks 199; Client Device 110; NL based Response System 120; Search System(s) 160).
Claims 6, 14 and 19,
Gray further teaches the method of claim 1, wherein: each of the first LLM and the second LLM are separate from, and in communication with, at least one device comprising the at least one processor ([Fig. 1] [0031] implemented remotely from the client device 110 (processor and memory));
the first LLM utilizes a first model algorithm to generate the first response ([0044-0045] LLM response generation engine using LLM to generate an NL based summary).
Inan further teaches the second LLM utilizes a second model algorithm to generate the second response ( [Abstract] Llama Guard, a Llama2-7b model, instruction-tuned).
Claims 7, 15 and 20,
Inan further teaches the method of claim 1, wherein the validity verdict indicates at least one inaccuracy criterion met by the first response ([3.1] if unsafe, output lists taxonomy categories (violated)).
Claim(s) 8 and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Gray et al. (US 2024/0220735) in view of Inan et al. (“Llama Guard: LLM based Input-Output Safeguard for Human-Ai Conversations”; Dec. 7, 2023) and further in view of Christian et al. (US 2023/0214892).
Claims 8 and 16,
Gray and Inan teach all the limitations in claim 1. The difference between the prior art and the claimed invention is that Gray nor Inan explicitly teach wherein the computing system comprises a tax calculation engine (TKE), and the processing state includes at least one of information received by the TKE from the UI, information received by the TKE from at least one additional source, a calculation performed by the TKE, tax data identified by the TKE as being relevant to the user, or a combination thereof.
Christian teaches wherein the computing system comprises a tax calculation engine (TKE), and the processing state includes at least one of information received by the TKE from the UI, information received by the TKE from at least one additional source, a calculation performed by the TKE, tax data identified by the TKE as being relevant to the user, or a combination thereof ([Figs. 9A-9B] [0034-0038] edge version of a tax calculation engine; receiving a tax calculation request from a client application; global tax rules database (tax rate and rule data); configured to calculate tax burden; identify a subset of the tax rate and rule data 28A applicable to each of the subset of products in each of the subset of geographic regions 54).
Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the teachings of Gray and Inan with teachings of Christian by modifying generative summaries for search results as taught by Gray to include wherein the computing system comprises a tax calculation engine (TKE), and the processing state includes at least one of information received by the TKE from the UI, information received by the TKE from at least one additional source, a calculation performed by the TKE, tax data identified by the TKE as being relevant to the user, or a combination thereof as taught by Christian for the benefit of calculating taxes applicable to transactions for goods and service at locations around the world (Christian [0002]).
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SHREYANS A PATEL whose telephone number is (571)270-0689. The examiner can normally be reached Monday-Friday 8am-5pm PST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Pierre Desir can be reached at 571-272-7799. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
SHREYANS A. PATEL
Primary Examiner
Art Unit 2653
/SHREYANS A PATEL/Examiner, Art Unit 2659
Prosecution Insights