Bank of America Corporation

Michael Young

Vinesh Patel

Manu Jacob Kurian

Ana Maxim

Kelly Renee-Drop Keiter

Kyle A Mayers

Sanjay Arjun Lohar

Robert Nyeland Huggins

Melissa Gordon Glenn

William R Overhultz, Jr

Kelly Keiter

Tarani Avasarala

Robert Hannon Sparks

Camilla Tyborowski

Graham Wyllie

Vinesh Young Patel

A method for providing a quantum reflection access credentialing (QRAC) system. The method includes receiving a user login. The user login may include a username and password. The user login may be associated with a known user. The system may include a memory for storing information corresponding to light reflected from the face of the user. The method may further include selecting, or receiving a selection of, a light source from among an array of light sources. The method may also include shining light generated by the selected light source on the user's face. The method may also include scanning the user face to capture a reflection of the light generated by the light source. The method may then include confirming that the information stored in the memory corresponds to the reflection of the light generated by the light source. In response to confirming that the information stored in the memory corresponds to the reflection of the light generated by the light source, the method may credential user access.

Quantum reflection access credentialing (QRAC)

Aspects of the disclosure relate an IoT authentication system to strengthen and streamline authentication for transaction processing for users. A transaction computing platform using artificial intelligence may receive from a plurality of IoT devices information regarding interactions with a user. In some arrangements, transaction computing platform may generate constellations based on the received information. A trust status may be determined based on the generated constellations. Transaction authentication may be based at least in part on the determined trust status. In some embodiments, real-time updates of trust status along with transaction alerts may be provided to IoT devices.

IoT based authentication

Apparatus and methods for an artificial intelligence implemented termination by an auditor artificial intelligence ('aAI') of a malicious artificial intelligence (“mAI”) are provided. The aAI may detect a mAI on a network and determine which data the mAI can access on the network. The aAI may then degrade all or part of the data in various ways to prevent the mAI from producing valid content based on the data. The aAI may also create code and inject the code into the mAI to degrade the mAI's operations. As the mAI may rely on valid data to produce valid output, degrading the data may degrade the mAI.

AI implemented AI termination

Apparatus, methods and systems for continual biometric authentication of a user on a user device is provided. Methods may create, for the user, a biometric authentication reference model. Methods may store the biometric authentication reference model at a database associated with the user device. In response to an instantiation of a session by the user on the user device, methods may execute a continual biometric scan. The scanning may be executed until the session is terminated. Methods may identify a number of comparison points between the scanned and stored biometric identifiers. In response to determining that the number of identified comparison points is below a predetermined threshold number of comparison points, methods may initiate a passcode request. The passcode request may request a confirmation of a preset user-selected passcode. In response to determining that the passcode response does not match the preset user-selected passcode, methods may terminate the session.

Continual biometric authentication

Apparatus, methods and systems for voice surveillance using enhanced metadata and geotagging capabilities are provided. Systems may monitor communications being transmitted from, and being received at, entity devices. The communications may be monitored for anomalies. The anomalies may be identified as communications outside of a regular pattern of communications. Various factors may be used to identify the anomalous communications. The factors may include a number of communications per predetermined time period, a geolocation of the communication, identification of a second communicative party of the communication and any other suitable factors. Stakeholders may be alerted regarding any anomalous communications. Additionally, once the communications have been labeled as anomalous or non-anomalous, the labeled communications may be used to create and/or update a set of labeled training data. The created and/or updated set of labeled training data may be used in an artificial intelligence engine that labels communications as anomalous or non-anomalous.

Voice surveillance system using enhanced metadata and geotagging capabilities

Systems and Method are provided for an Anti-Pattern Captcha. Methods may include AI software receiving a request from a user to access a system or website, the request comprising user credentials. Methods may include the AI software identifying a human user associated with the user credentials. Methods may include the AI software compiling data related to a plurality of unique behavior patterns associated with the human user when interacting online. Methods may include the AI software analyzing the compiled data to establish the unique behavior patterns. Methods may include the AI software presenting a challenge to the user. Methods may include the AI software verifying that the user is the human user upon a response to the challenge that deviates from at least one of the unique behavior patterns. Methods may include the AI software prompting the user to interact with the system or website upon verification.

Anti-pattern captcha

Artificial intelligence (AI) impersonation detection using an AI model is provided. Methods may train an AI model. The training may provide the AI model with a first dataset including a communication between two or more human users, a second and third dataset including an impersonation of the communication by a public AI model and a private AI model respectively. Methods may identify a first identifier set in the first dataset, a second identifier set in the second dataset and a third identifier set in the third dataset. Methods may create a parameter range for the first, second and third identifier set. The methods may include monitoring a production communication using the AI model. The AI model may compare the production communication with the parameter ranges. Methods may identify a smallest comparison value. Methods may, based on the identified smallest comparison value, identify the origin of the production communication.

Artificial intelligence impersonation detector

Systems and methods for an artificial intelligence-based auditor of artificial intelligence may be provided. The artificial-intelligence-based auditor may operate on one or more hardware processors. The artificial intelligence-based auditor may continually scan one or more operating artificial intelligence systems for productivity and operability. The scan may identify a data source powering the one or more operating artificial intelligence systems. The scan may measure a delta between an output from each of the one or more operating artificial intelligence systems. The scan may identify trends in quality of the one or more operating artificial intelligence systems. The scan may analyze instances where the one or more operating artificial intelligence systems outputted data outside of a parameter range. Based on the scan, the auditor may label each of the one or more operating artificial intelligence systems as positive, neutral or negative.

Artificial intelligence-based auditors of artificial intelligence

Apparatus and methods for using deepfakes defensively to detect fake, spoofed, and hoax accounts and posts on social media and elsewhere are provided. A program may gather verified images or writings of a target individual. The program may analyze the verified images and writings to create a baseline. The program may use deepfake algorithms to create exemplar deepfake images or writings. The program may store the data in a database. The program may search a network for social media accounts or posts that may meet the baseline and determine whether they are legitimate or illegitimate by comparing the contents of the accounts or posts with the exemplar deepfakes. When the program determines the accounts or posts are illegitimate, the program may initiate a takedown of those accounts or posts. The program may use machine learning algorithms to refine itself and become more accurate.

Defensive deepfake for detecting spoofed accounts

Methods for securing an electronic communication is provided. In a registration process, an anti-phish, personalized, security token may be created and/or selected for a predetermined account. The token may be stored in a database at an enterprise location. An electronic communication may be generated at a third-party location on behalf of the enterprise. The communication may be forwarded from the third-party location to a recipient associated with the account. The communication may be intercepted at an edge server. The edge server may be located at the third-party location or the enterprise location. The edge server may be in communication with the database. The edge server may select, from the database, the anti-phish token that is associated with the account. The selected token may be injected into the communication. The communication with the token may be transmitted to the recipient.

Growing community of inventors

Davidson, NC, United States of America

Michael Young