Using AI to Identify and Prevent Fraud
Since the beginning, Cash App has implemented safeguards to prevent and combat illicit activity on the platform. AI and machine learning technology can help us quickly and precisely identify and thwart illegal and fraudulent activities. AI is a critical tool in helping us fight fraud and criminal activity in real time and before it even happens.
Keeping Accounts Safe in Real Time: AI and Transaction Tags
Cash App uses advanced security features to prevent unauthorized users from accessing customer accounts. We use several layers of tools — both automated and human-prompted — to detect account takeovers and lock fraudsters out while maintaining a quick review period for all locked accounts to ensure good customers regain access swiftly. AI-powered automated tools are often the first line of defense to identify and prevent malicious activity.
In some cases, our machine learning models can identify and prevent fraud before a transaction is even attempted. In others, when we determine an account was accessed without authorization, we can reset certain account attributes (such as phone number) that could have been modified by a fraudster. For example, if a fraudster added their phone number or email so that they could log back into the account, resetting those characteristics would prevent future unauthorized access via those attributes.
In addition to manually identifying and sweeping for suspicious keywords, our AI-powered models provide efficient, effective, and adaptable tools to identify and lock out fraudsters. When customers provide us with their feedback, their notes provide a valuable piece of information for our models to consider in identifying which transactions were unauthorized. Using natural language processing, our models can adapt to language in these notes and secure accounts quickly and accurately.
To read a short case study on how Cash App has used natural language processing to protect customers, visit the Cash App AI Blog: Using Natural Language Processing to Reduce Account Takeovers.
Since the beginning, Cash App has implemented safeguards to prevent and combat illicit activity on the platform. AI and machine learning technology can help us quickly and precisely identify and thwart illegal and fraudulent activities. AI is a critical tool in helping us fight fraud and criminal activity in real time and before it even happens.
Keeping Accounts Safe in Real Time: AI and Transaction Tags
Cash App uses advanced security features to prevent unauthorized users from accessing customer accounts. We use several layers of tools — both automated and human-prompted — to detect account takeovers and lock fraudsters out while maintaining a quick review period for all locked accounts to ensure good customers regain access swiftly. AI-powered automated tools are often the first line of defense to identify and prevent malicious activity.
In some cases, our machine learning models can identify and prevent fraud before a transaction is even attempted. In others, when we determine an account was accessed without authorization, we can reset certain account attributes (such as phone number) that could have been modified by a fraudster. For example, if a fraudster added their phone number or email so that they could log back into the account, resetting those characteristics would prevent future unauthorized access via those attributes.
In addition to manually identifying and sweeping for suspicious keywords, our AI-powered models provide efficient, effective, and adaptable tools to identify and lock out fraudsters. When customers provide us with their feedback, their notes provide a valuable piece of information for our models to consider in identifying which transactions were unauthorized. Using natural language processing, our models can adapt to language in these notes and secure accounts quickly and accurately.
To read a short case study on how Cash App has used natural language processing to protect customers, visit the Cash App AI Blog: Using Natural Language Processing to Reduce Account Takeovers.