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 Amendments
This communication is in response to the amendments filed on 25 June 2026:
Claims 1, 11 and 20 are amended.
Claims 1-20 are pending.
Response to Arguments
In response to Applicant’s remarks filed on 25 June 2026:
a. Applicant’s arguments that Stapleton and Canedo do not teach or suggest “the response indicator suggesting an alternative response, which is configured to replace a portion of an AI response that is received from the AI model as a result of the AI prompt being processed with a replacement portion, as a response to the AI prompt” regarding claims 1-19 has been fully considered and is deemed fully persuasive. Applicant’s attention is directed to the allowable subject matter presented below.
b. Applicant’s arguments that Stapleton and Canedo do not teach or suggest “as a result of the response indicator suggesting the alternative response in lieu of the AI response as the response to the AI prompt, provide the alternative response in lieu of the AI response as the response to the AI prompt” has been fully considered and is deemed fully persuasive with respect to claims 1-19, which require the alternative response comprising a replacement portion for an AI response that is received from the AI model. As for claim 20, the arguments have been fully considered but are deemed not-persuasive. Applicant’s attention is directed to Stapleton, Paragraph [0093], see “…the system may cause one or more computing devices to provide output that indicates that the one or more parameters of the trained machine learning model have been compromised…”, which is being read as providing the alternative response (e.g., model has been compromised) in lieu of the AI response as the response to the AI prompt. The indication that the model has been compromised (e.g., alternative response) is responsive to the AI prompt, whether the AI prompt had malicious input or was a directed prompt injection. The Examiner suggests amending claim 20 to include the limitation of the alternative response being configured to replace a portion of the AI response that is received from the AI model as a result of the AI prompt being processed with a replacement portion, as a response to the AI prompt.
c. Applicant’s arguments that Stapleton and Canedo do not teach or suggest “the response indicator suggesting an alternative response that is responsive to the AI prompt in lieu of the an AI response, which is received from the AI model as a result of the AI prompt being processed, as a response to the AI prompt” regarding claim 20 has been fully considered but is deemed not-persuasive. Applicant’s attention is directed to Stapleton, Paragraph [0093], see “…the system may analyze the decrypted output data to determine whether one or more of the parameters of the trained machine learning model have been compromised…the analysis may include determining whether the decrypted output data complies with an expected output structure…If these constraints are not satisfied by decrypted data, that may indicate that the model has been compromised…the system may cause one or more computing devices to provide output that indicates that the one or more parameters of the trained machine learning model have been compromised…”, where the response indicator suggests an alternative response in lieu of an AI response (e.g., providing an output that indicates that the model has been compromised).
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 20 is rejected under 35 U.S.C. 103 as being unpatentable over STAPLETON et al. (WO 2020/151964), hereinafter Stapleton, in view of Canedo et al. (U.S. PGPub. 2021/0150359), hereinafter Canedo.
Regarding claim 20, Stapleton teaches A computing system comprising:
a memory that stores an operating system and an artificial intelligence (AI) model (Stapleton, Paragraph [0013], see “…execute instructions stored in associated memory…”) (Stapleton, Paragraph [0035], see “…one or more ML models may be stored by AI provider system 100 in a ML model database 104…”) (Stapleton, Paragraph [0040], see “…a ML model is stored remotely from AI provider system 100, e.g., in database 116, one or more clients…may host a software application that is operable by end users 114 to make use of the ML model…”), the operating system including a utility that is configured to transfer encrypted communications between a (Stapleton, Fig. 1, see “102”, “110”, and “100”, wherein the utility (the connection between 102 and 110) is configured to transfer encrypted communications between a processing unit (comprised in 102) and a cloud-based security service (“AI PROVIDER SYSTEM 100”)); and
the
execute the AI model (Stapleton, Paragraph [0038], see “…Application engine 107 may apply this input data across one or more ML models…to generate output”, which is being read as executing the AI model);
encrypt AI interaction data (Stapleton, Paragraph [0008], see “…applying the encrypted input data as input across the encrypted training machine learning model to generate encrypted output”, where “encrypted output” is being read as comprising encrypted AI interaction data), which includes an AI prompt (Stapleton, Paragraph [0008], see “…applying the encrypted input data as input across the encrypted training machine learning model…”, where “encrypted input data” is being read as comprising an AI prompt), using a cryptographic key to provide encrypted AI interaction data (Stapleton, Paragraph [0057], see “Before or during input stage 438, an encryption key 446 may be provided, e.g., by AI provider system 100 to one or more remote computing systems 102…This encryption key 446 may be used by one or more users to generate, from data provided by sources 444, encrypted data 448. When the time comes to apply the encrypted input data 448 across one or more ML models…various actions may be taken”, where “encryption key” is being read as a cryptographic key which is used to provide the encrypted AI interaction data);
provide the encrypted AI interaction data to the cloud-based security service (Stapleton, Paragraph [0033], see “…AI provider system 100 may provide, to one or more individuals access to one or more machine learning (“ML”) models”, wherein the AI models (machine learning models) are stored locally on the cloud-based security service (e.g., AI provider system), hence the input/output data is provided to the AI provider system) via the utility in the operating system (Stapleton, Paragraph [0032], see “…AI provider system 100 may be communicatively coupled with one or more remote computing systems…over one or more wired and/or wireless computing networks”);
receive a response indicator from the cloud-based security service via the utility in the operating system (Stapleton, Paragraph [0093], see “…the system may analyze the decrypted output data to determine whether one or more of the parameters of the trained machine learning model have been compromised”), the response indicator representing a result of an analysis of a decrypted representation of the encrypted AI interaction data, the analysis including at least one of a security analysis or a sensitivity analysis, the response indicator suggesting an alternative response that is responsive to the AI prompt in lieu of an AI response, which is received from the AI model as a result of the AI prompt being processed, as a response to the AI prompt (Stapleton, Paragraph [0093], see “…the system may analyze the decrypted output data to determine whether one or more of the parameters of the trained machine learning model have been compromised…the analysis may include determining whether the decrypted output data complies with an expected output structure…If these constraints are not satisfied by decrypted data, that may indicate that the model has been compromised…the system may cause one or more computing devices to provide output that indicates that the one or more parameters of the trained machine learning model have been compromised…”, where “analyze the decrypted output data” is being read as indicating a result of an analysis of a decrypted representation of the encrypted AI interaction data, the analysis including at least one of a security analysis, and the response indicator suggesting an alternative response in lieu of an AI response (e.g., providing an output that indicates that the model has been compromised)); and
as a result of the response indicator suggesting the alternative response in lieu of the AI response as the response to the AI prompt, provide the alternative response in lieu of the AI response as the response to the AI prompt (Stapleton, Paragraph [0093], see “…the system may cause one or more computing devices to provide output that indicates that the one or more parameters of the trained machine learning model have been compromised…”, which is being read as providing the alternative response (e.g., model has been compromised) in lieu of the AI response as the response to the AI prompt).
Stapleton does not teach the following limitation(s) as taught by Canedo: the operating system including a utility that is configured to transfer encrypted communications between a neural processing unit and a cloud-based security service (Canedo, FIG. 2, where encrypted communications are transferred between a neural processing unit and a cloud-based security service) (Canedo, Paragraph [0033], see “…the training and testing process for neural networks can be performed either in the host system (e.g., laptop or the cloud) or in the neural co-processor(s)…”).
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the techniques disclosed of Stapleton, by implementing techniques of communications between a neural processor and a cloud service, disclosed of Canedo.
One of ordinary skill in the art would have been motivated to make this modification in order to implement techniques for secure execution of an AI model of a neural processing unit, comprising of communications between a neural processor and a cloud service. This allows for a more optimal performance and improved model accuracy by combining real-time capabilities of the neural processor with the virtually limitless resources of the cloud service. Canedo is deemed as analogous art due to the art disclosing techniques of communications between a neural processor and a cloud service (Canedo, FIG. 2).
Allowable Subject Matter
Claims 1-19 are allowed.
Conclusion
THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 extension fee 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 date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RODMAN ALEXANDER MAHMOUDI whose telephone number is (571)272-8747. The examiner can normally be reached on M-F 11:00am – 7:00pm.
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, Philip Chea can be reached on (571) 272-3951. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/RODMAN ALEXANDER MAHMOUDI/Examiner, Art Unit 2499
Prosecution Insights