Bank of America Corporation

Sanjay Arjun Lohar

James J Siekman

Kyle A Mayers

Kelly Renee-Drop Keiter

Michael James Sbandi

Michael Robert Young

Dustin Paul Stocks

Andrew Thomas Boyd

Daniel Mercado

Maharaj Mukherjee

George Albero

Gilbert M Gatchalian

Olga Kocharyan

Vijaya L Vemireddy

Richard Scot

Cody Dean Searl

Michael Young

Erik Dahl

Benjamin F Tweel

Melissa Gordon Glenn

Karen E Davis

Dinesh Kumar Agrawal

Ramkumar Korlepara

Steven Nathan Greene

Eric Eugene Sifford

David Aloysius Cuka

Lisa Snow Pugh

Kelly Keiter

W Scott Hammet

James R Crocker

Nicholas Stephen Rose

Connor Thomas McCormick

Charlene Moss Gorter

Marabella Youkhaneh White

Pamela Elizabeth Barnes

Marisa Kamer

Anna Maria Mirarchi

Margaret M Brewer

Anna Xu

Samantha Ilse Quales

Anna E Ganse

Kayla Ashley Rux

Viswanathan Venkatasubramanian

Ashley L Jones

A method for analyzing messages for malicious content is provided. The method includes intercepting a message comprising a quick response (QR) code or a uniform resource locator (URL). The QR code corresponds to the URL. A status of the message is determined by determining a source of the message. The source of the message is compared to known malicious sources. In response to the source of the message not matching any of the known malicious sources, the source of the message is compared to known safe sources. In response to the source of the message not matching any of the known safe sources, the URL is compared to known malicious URLs. In response to the URL matching a respective one of the known malicious URLs, the status of the message is determined as malicious. The message is released to be displayed along with the status.

Analyzing messages for malicious content using a cloud computing system

A system is provided for authentication of resource transfers using tokenization as indicator of authorized resource distribution. In particular, the system may be configured to generate a secure transferor token on a distributed register, where each transferor token may be associated with a resource transferor within a virtual environment. The transferor token may be digitally signed using a transferor private key such that the transferor may be authenticated by entities using the corresponding public key. The system may further generate a secure token associated with the resource to be transferred, where the token associated with the resource may specify a list of authorized transferors as one of its parameters. Accordingly, the system may ensure that resources are transferred by authorized transferors.

System and method for authentication of resource transfers using tokenization as indicator of authorized resource distribution

A device for performing a secured operation is disclosed. The device is embedded with a predefined data item and verification information. The verification information is uniquely associated with the user of the device. The verification information is linked to the identity of the user. The verification information is configured to be extracted in conjunction with the secured operation being performed. The secured operation is performed in response to the verification information being verified. A system and method for performing the secured operation are also disclosed.

System, device, and method for performing secured operations

Apparatus and methods for using deepfakes defensively to detect fake, spoofed, and hoax phone calls and videoconferences are provided. A program may record a target individual reciting exemplary phrases. The program may analyze the recordings to create a baseline. The program may use deepfake algorithms to create exemplar deepfake audiovisual representations of the target individual. The program may store the data in a database. The program may analyze, in real-time, a phone call or videoconference to determine whether they are legitimate or illegitimate by comparing the audio or audiovisual contents of the phone call or videoconference with the exemplar deepfakes. When the program determines the phone call or videoconference is illegitimate, the program may terminate the call or videoconference and inform the recipient and others.

Defensive deepfake for detecting live deepfaked audio and video

Arrangements for detecting unauthorized activity based on input method analysis and monitoring are provided. In some aspects, identity information associated with a user may be received and be stored. An input may be received from a computing device of the user. An input pattern of the received input may be determined. Using a machine learning model, the input pattern of the received input may be compared to input patterns of humans and input patterns of machines. Based on the comparison, it may be determined whether the user is a human user or a non-human user. Responsive to determining that the user is a non-human user, a request may be transmitted to the user to provide increased authentication credentials. Responsive to determining that the user is a human user, an identity of the user may be verified by comparing the input pattern of the received input to the stored identity information.

Unauthorized activity detection based on input analysis and monitoring

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

Arrangements for providing software vulnerability analysis and monitoring are provided. In some aspects, software bill of materials (SBOM) data may be received and software attributes may be extracted from the SBOM data. Author data may be received and analyzed using natural language processing and/or machine learning to identify author attributes. Current event or vulnerability data may be received. In some examples, one or more machine learning models may be executed to determine a confidence score associated with the software being analyzed. For instance, software attributes, author attributes, and current event data may be used as inputs in the machine learning model and a confidence score may be output. Based on the confidence score, one or more alerts may be generated and transmitted to one or more enterprise organization computing devices.

Multi-computer system for performing vulnerability analysis and alert generation

Apparatus and methods for providing an indicator of trustworthiness of a software application are provided. A program may receive information regarding the software application from a developer of the application. The program may analyze the application and determine that the application is trustworthy and secure. The program may then create a unique logo that may include various information to indicate to a user that the program has determined the application is trustworthy. The program may transmit the logo to the developer so the developer may incorporate the logo into the application. In an embodiment, the program may transmit the logo to a third-party application store for display on a page where the application is available.

Logos as visual indicators of trust in applications

A system includes a plurality of sensors distributed about a space. Each sensor generates sensor data associated with properties of one or more people in the space. A gateway device receives the sensor data from each of the plurality of sensors and provides the sensor data to an automated response subsystem over a network. The automated response subsystem receives the sensor data. A first feature value is determined based on the received sensor data. An event score is determined based on the first feature value. The event score is compared to predefined response criteria. Based on the comparison of the event score to the predefined response criteria, an actionable event is detected, and an automated response to initiate for the detected actionable event is identified. One or both of an alert and instructions are provided in order to initiate the automated response.

System for detecting unauthorized activity

A system includes a plurality of sensors distributed about a space. Each sensor generates sensor data associated with properties of one or more people in the space. An automated response subsystem receives the sensor data. An unauthorized activity performed by a first person in the space is detected based on at least a portion of the received sensor data. In response to detecting the unauthorized activity, tracking instructions are transmitted that identify the first person associated with the unauthorized activity. A tracking device receives the tracking instructions. A device associated with the first person is detected. Over a period of time, the tracking device sends queries to the device and receives corresponding responses from the device. The tracking device determines positions of the device within the space over the period of time.

System for detecting and tracking an unauthorized person

A plurality of communicatively coupled, networked assets may be threatened or attacked by a cybersecurity attack. The operational resiliency of the computer network determines whether the cybersecurity attack leads to a shutdown of one or more assets, or even the entire computer network. Machines and processes are disclosed to improve operational cybersecurity resiliency of software on the computer network. Machine learning is used to identify potential vulnerabilities from a vulnerability database. Chaos stress testing using a machine learning algorithm can be performed on software to exploit the vulnerabilities. A blast radius can be set to minimize any potential negative side effects of the testing. Software can be remediated to account for responses to the testing by reconfiguring to prevent exploitation of the vulnerabilities. A financial impact of the exploited vulnerabilities can be calculated and reports can be generated.

Cyber resilience chaos stress testing

Growing community of inventors

Charlotte, NC, United States of America

Sanjay Arjun Lohar