Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
2. The information disclosure statements (IDS) submitted on March 02, 2023, July 31, 2023, and July 29, 2024 are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner
Response to Amendments and Arguments
3. The amendment filed on July 01, 2025 has been entered. Claims 1-16 remain pending in the application. Claim 1 has been amended.
The applicant argues that Collazo in view of Carbune does not disclose or suggest integration of stylometric analysis with typing-based behavioral signals for the purpose of authorship determination. However, the examiner respectfully disagrees with this assertion. Support can be found in Collazo and Carbune. Paragraphs [0050 and [0053] of Collazo disclose the following: [0050] - a user behavior session, such as user behavior session 210 and user behavior session 214, defines the session where characteristics of functional user behavior on an application are collected. A user behavior session may be associated with a particular browser cookie. The information that security system 200 keeps track of during a user behavior session may include a web page or screen sequence navigated by the user, along with the time intervals associated with each page. Other information about functional user behavior may include input in web pages, screens, or forms of the application. Other examples may include timestamped user clicks or touch strokes within a web page or screen of the application… [0053] - The behavior tracker script and behavior analysis engine 216 may be configured to track and store a predefined set of various kinds of functional user behaviors. These paragraphs disclose the tracking and storing of user behavior through tracking inputs, including timing data – “timestamped user clicks or touch strokes within a web page or screen of the application…”, which is all found in claim 1. There is not mention of full-text drafting sessions, such as recording revision patterns, word choice evolution in the independent claims. Furthermore, paragraph [0044] of Carbune discloses the following: “The discriminator 210 then determines, for each dataset, a prediction 230 of whether a given input is computer-generated (i.e., synthetic) or human-generated (i.e., real))”. Carbune discloses determining whether input is text written by human or artificial intelligence. There is no mention of nuanced indicators of human authorship such as word substitutions, spelling variations, editing patterns, and structural revision within extended writing in the independent claims. Finally, the two references can be understood to combined using motivation found in paragraph [0054] of Carbune, which states that the Carbune reference can be used to improve determination whether responses are human or computer.
Hence, the applicant’s arguments are not persuasive.
Claim Rejections - 35 USC § 103
3. 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 taught 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.
4. Claims 1-8 and 10-16 are rejected under 35 U.S.C. 103 as being unpatentable over Collazo (U.S. Publication No. 20100299292) in view of Carbune (U.S. Publication No. 20200342879).
Regarding claim 1, Collazo discloses a method for determining authorship of a text is written by a human or an artificial intelligence, comprising:
tracking inputs to a computer during generation of the text ([0053] - a behavior tracker script (e.g., JavaScript file) is installed on pages of web application 222 to track user clicks or text input or any functional user behavior);
storing metadata from the tracked inputs, said metadata including time of entry of inputs and patterns of entry of the inputs ([0050] - a user behavior session, such as user behavior session 210 and user behavior session 214, defines the session where characteristics of functional user behavior on an application are collected. A user behavior session may be associated with a particular browser cookie. The information that security system 200 keeps track of during a user behavior session may include a web page or screen sequence navigated by the user, along with the time intervals associated with each page. Other information about functional user behavior may include input in web pages, screens, or forms of the application. Other examples may include timestamped user clicks or touch strokes within a web page or screen of the application… [0053] - The behavior tracker script and behavior analysis engine 216 may be configured to track and store a predefined set of various kinds of functional user behaviors);
comparing the metadata to statistical values for time of entry and patterns of entry during composition in human-generated text and in artificial intelligence generated text ([0060] - Correlation engine 218 correlates current behavior patterns with known evaluation patterns. For instance, information about page sequence in a user behavior session stored in relational database 226 may be transformed into a state in a Markov chain; the state may be used as part of the correlation process to determine whether the current state is associated with malicious behavior or not);
However, Collazo does not disclose on the basis of the comparison, issuing a determination whether the text was written by a human or by an artificial intelligence.
Carbune does teach on the basis of the comparing, issuing a determination whether the text was written by a human or by an artificial intelligence ([0044] - The discriminator 210 then determines, for each dataset, a prediction 230 of whether a given input is computer-generated (i.e., synthetic) or human-generated (i.e., real)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement on the basis of the comparing, issuing a determination whether the text was written by a human or by an artificial intelligence. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
Regarding claim 2, Collazo in view of Carbune teaches all limtations of claim 1, above.
However, Collazo does not disclose the method, further comprising generating a certificate attesting to human or artificial intelligence authorship.
Carbune does teach the method, further comprising generating a certificate attesting to human or artificial intelligence authorship ([0074] - the decision engine 150 may flag a remote entity 160 as malicious based on receiving a prediction from the A-Discriminator 134 that the remote entity 160 is not human. Similarly, the decision engine 150 may flag a remote entity 160 as a trusted entity based upon receiving a prediction from the A-Discriminator 134 that the remote entity 160 is human, as expected).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement the method, further comprising generating a certificate attesting to human or artificial intelligence authorship. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
Regarding claim 3, Collazo in view of Carbune teaches all limtations of claim 2, above.
However, Collazo does not disclose the method, wherein the certificate is a digital signature that is appended to the document.
Carbune does teach the method, wherein the certificate is a digital signature that is appended to the document ([0074] - the decision engine 150 may flag a remote entity 160 as malicious based on receiving a prediction from the A-Discriminator 134 that the remote entity 160 is not human. Similarly, the decision engine 150 may flag a remote entity 160 as a trusted entity based upon receiving a prediction from the A-Discriminator 134 that the remote entity 160 is human, as expected).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement the method, wherein the certificate is a digital signature that is appended to the document. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
Regarding claim 4, Collazo in view of Carbune teaches all limtations of claim 2, above.
However, Collazo does not disclose the method, wherein the certificate comprises a reference within the text to an external document containing an analysis of the metadata.
Carbune does teach the method, wherein the certificate comprises a reference within the text to an external document containing an analysis of the metadata ([0074] - the decision engine 150 may flag a remote entity 160 as malicious based on receiving a prediction from the A-Discriminator 134 that the remote entity 160 is not human. Similarly, the decision engine 150 may flag a remote entity 160 as a trusted entity based upon receiving a prediction from the A-Discriminator 134 that the remote entity 160 is human, as expected).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement the method, wherein the certificate comprises a reference within the text to an external document containing an analysis of the metadata. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
Regarding claim 5, Collazo in view of Carbune teaches all limtations of claim 1, above Collazo discloses the method, wherein the comparing step comprises performing a specific comparison for a plurality of units of the text, and wherein the step of issuing a determination comprises issuing a specific determination for each of the plurality of units ([0026] - The behavior pattern is correlated with each of the set of evaluation patterns. In some embodiments, the behavior pattern continues to be correlated with evaluation patterns as more information about application-level functional user behavior is collected. Correlation methods may comprise checking the quantitative values in the behavior pattern against some known value, such as a threshold).
Regarding claim 6, Collazo in view of Carbune teaches all limtations of claim 1, above.
However, Collazo does not disclose the method, wherein the text is categorized based on at least one of a category of composition or a category of author, and the comparing step comprises determining whether the text was written by a human or by an artificial intelligence based on statistical values corresponding to said category of composition or author.
Carbune does teach the method, wherein the text is categorized based on at least one of a category of composition or a category of author, and the comparing step comprises determining whether the text was written by a human or by an artificial intelligence based on statistical values corresponding to said category of composition or author ([0040] - The entity tagger of the natural language processor 126 may annotate references to an entity at a high level of granularity (e.g., to enable identification of all references to an entity class such as people) and/or a lower level of granularity (e.g., to enable identification of all references to a particular entity such as a particular person). The entity tagger may rely on content of the natural language input to resolve a particular entity and/or may optionally communicate with a knowledge graph or other entity database to resolve a particular entity).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement the method, wherein the text is categorized based on at least one of a category of composition or a category of author, and the comparing step comprises determining whether the text was written by a human or by an artificial intelligence based on statistical values corresponding to said category of composition or author. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
Regarding claim 7, Collazo in view of Carbune teaches all limtations of claim 1, above.
However, Collazo does not disclose the method, wherein the comparing step comprises determining whether the text was written by a specific person.
Carbune does teach the method, wherein the comparing step comprises determining whether the text was written by a specific person ([0040] - The entity tagger of the natural language processor 126 may annotate references to an entity at a high level of granularity (e.g., to enable identification of all references to an entity class such as people) and/or a lower level of granularity (e.g., to enable identification of all references to a particular entity such as a particular person). The entity tagger may rely on content of the natural language input to resolve a particular entity and/or may optionally communicate with a knowledge graph or other entity database to resolve a particular entity).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement the method, wherein the comparing step comprises determining whether the text was written by a specific person. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
Regarding claim 8, Collazo in view of Carbune teaches all limtations of claim 1, above.
Collazo discloses the method, further comprising separately evaluating language patterns in the text, and the step of issuing a determination further comprises issuing a combined determination based on both valuation of the language patterns and evaluation of the metadata ([0050] - a user behavior session, such as user behavior session 210 and user behavior session 214, defines the session where characteristics of functional user behavior on an application are collected. A user behavior session may be associated with a particular browser cookie. The information that security system 200 keeps track of during a user behavior session may include a web page or screen sequence navigated by the user, along with the time intervals associated with each page. Other information about functional user behavior may include input in web pages, screens, or forms of the application. Other examples may include timestamped user clicks or touch strokes within a web page or screen of the application… [0053] - The behavior tracker script and behavior analysis engine 216 may be configured to track and store a predefined set of various kinds of functional user behaviors).
Regarding claim 10, Collazo in view of Carbune teaches all limitations of claim 10, above.
Collazo discloses the method, wherein the inputs comprise one or more of: keystores, cursor movments, mouse clicks, and mouse location ([0050] - The information that security system 200 keeps track of during a user behavior session may include a web page or screen sequence navigated by the user, along with the time intervals associated with each page. Other information about functional user behavior may include input in web pages, screens, or forms of the application. Other examples may include timestamped user clicks or touch strokes within a web page or screen of the application. The functional user behavior may be processed and tagged by the context of the behavior. For instance, a click on an image on a web page may be tagged with a description of the image (e.g., "information icon")).
Regarding claim 11, Collazo in view of Carbune teaches all limitations of claim 1, above.
Collazo discloses the method, wherein the inputs comprise use of word processing editing functions ([0053] - The data collected may be aggregated, processed, transformed and analyzed by behavior analysis engine 216. In some embodiments, information about functional user behavior is transformed from raw data collected from users and/or clients into behavior patterns, where the behavior pattern is in a format that matches with the inputs of the correlation methods. The behavior patterns stored and collected in relational database 226 may be used by learning engine 220 to improve the algorithms running on behavior analysis engine 216 and correlation engine 218).
Regarding claim 12, Collazo in view of Carbune teaches all limitations of claim 1, above.
However, Collazo does not disclose the method, wherein the inputs comprise languages used during the inputting.
Carbune does teach the method, the method, wherein the inputs comprise languages used during the inputting ([0039] - the natural language processor 126 is configured to identify and annotate various types of grammatical information in natural language input. For example, the natural language processor 126 may include a part of speech tagger (not depicted) configured to annotate terms with their grammatical roles. Also, for example, in some implementations the natural language processor 126 may additionally and/or alternatively include a dependency parser (not depicted) configured to determine syntactic relationships between terms in natural language input).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement the method, wherein the inputs comprise languages used during the inputting. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
Regarding claim 13, Collazo in view of Carbune teaches all limitations of claim 1, above.
Collazo discloses the method, wherein the time of entry of inputs comprises both time spent for entry of specific units of text and cumulative time of entry of the text ([0050] - The information that security system 200 keeps track of during a user behavior session may include a web page or screen sequence navigated by the user, along with the time intervals associated with each page. Other information about functional user behavior may include input in web pages, screens, or forms of the application. Other examples may include timestamped user clicks or touch strokes within a web page or screen of the application. The functional user behavior may be processed and tagged by the context of the behavior. For instance, a click on an image on a web page may be tagged with a description of the image (e.g., "information icon")).
Regarding claim 14, Collazo in view of Carbune teaches all limitations of claim 1, above.
Collazo discloses the method, wherein the patterns of entry of inputs comprise patterns of revision to the text ([0062] - The new evaluation patterns may represent new user behavior patterns, perhaps due in part to changes in the application or changes in user activities on the application)
Regarding claim 15, Collazo in view of Carbune teaches all limitations of claim 1, above.
However, Collazo does not disclose the method, further comprising determining whether to publish a text on a basis of the determination of whether the text was generated by a human or by an artificial intelligence.
Carbune does teach the method, the method, further comprising determining whether to publish a text on a basis of the determination of whether the text was generated by a human or by an artificial intelligence ([0047] - a human may interact with an automated assistant (e.g., 110) and the automated assistant may generate a dialog with the human. The automated assistant can provide prompts, such as questions, to the human, and the human may respond to the questions. The human answers may then be stored in the human answer database 135 to be utilized as training data for the A-GAN 130).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement the method, further comprising determining whether to publish a text on a basis of the determination of whether the text was generated by a human or by an artificial intelligence. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
Regarding claim 16, Collazo discloses a computer program product comprising instructions stored on a non-transitory computer-readable medium that, when executed by a computer, causes performance of the following steps:
tracking inputs to a computer during generation of the text ([0053] - a behavior tracker script (e.g., JavaScript file) is installed on pages of web application 222 to track user clicks or text input or any functional user behavior);
storing metadata from the tracked inputs, said metadata including time of entry of inputs and patterns of entry of the inputs ([0050] - a user behavior session, such as user behavior session 210 and user behavior session 214, defines the session where characteristics of functional user behavior on an application are collected. A user behavior session may be associated with a particular browser cookie. The information that security system 200 keeps track of during a user behavior session may include a web page or screen sequence navigated by the user, along with the time intervals associated with each page. Other information about functional user behavior may include input in web pages, screens, or forms of the application. Other examples may include timestamped user clicks or touch strokes within a web page or screen of the application… [0053] - The behavior tracker script and behavior analysis engine 216 may be configured to track and store a predefined set of various kinds of functional user behaviors);
comparing the metadata to statistical values for time of entry and patterns of entry during composition in human-generated text and in artificial intelligence generated text ([0060] - Correlation engine 218 correlates current behavior patterns with known evaluation patterns. For instance, information about page sequence in a user behavior session stored in relational database 226 may be transformed into a state in a Markov chain; the state may be used as part of the correlation process to determine whether the current state is associated with malicious behavior or not);
However, Collazo does not disclose on the basis of the comparing, issuing a determination whether the text was written by a human or by an artificial intelligence.
Carbune does teach on the basis of the comparing, issuing a determination whether the text was written by a human or by an artificial intelligence ([0044] - The discriminator 210 then determines, for each dataset, a prediction 230 of whether a given input is computer-generated (i.e., synthetic) or human-generated (i.e., real)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo to incorporate the teachings of Carbune in order to implement on the basis of the comparing, issuing a determination whether the text was written by a human or by an artificial intelligence. Doing so allows improved determination whether responses are human or computer (Carbune [0054]).
5. Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Collazo (U.S. Publication No. 20100299292) in view of Carbune (U.S. Publication No. 20200342879) in view of Tran (U.S. Publication No. 20220237368).
Regarding claim 9, Collazo in view of Carbune teaches all limitations of claim 1, above.
However, Collazo in view of Carbune does not teach the method, further comprising evaluating language patterns in the text, comparing the language patterns to published texts, and issuing a determination regarding whether any portion of the text was plagiarized.
Tran does teach the method, further comprising evaluating language patterns in the text, comparing the language patterns to published texts, and issuing a determination regarding whether any portion of the text was plagiarized ([0022] - detecting plagiarism in the document by matching the document text to text crawled from the Internet).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Collazo in view of Carbune to incorporate the teachings of Tran in order to implement the method, further comprising evaluating language patterns in the text, comparing the language patterns to published texts, and issuing a determination regarding whether any portion of the text was plagiarized. Doing so allows for significant speed in document generation, while cost is reduced (Tran [0036]).
Conclusion
6. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Cheng (U.S. Patent No. 11481678) teaches systems and methods for learning new watermark algorithms for a data processing accelerator. Galle (U.S. Publication No. 20230109734) teaches computer-implemented method for distributional detection of machine-generated text. Haikin (U.S. Patent No. 11984116) teaches method and system for unsupervised discovery of unigrams in speech recognition systems. Rath (U.S. Patent No. 11468232) teaches detecting machine text. Takaki (U.S. Patent No. 9372850) teaches machined book detection.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, 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 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 ETHAN DANIEL KIM whose telephone number is (571) 272-1405. The examiner can normally be reached on Monday - Friday 9:00 - 5:00.
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, Richemond Dorvil can be reached on (571) 272-7602. 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 https://ppair-my.uspto.gov/pair/PrivatePair. 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.
/ETHAN DANIEL KIM/
Examiner, Art Unit 2658
/RICHEMOND DORVIL/Supervisory Patent Examiner, Art Unit 2658