In a significant move to strengthen its fraud prevention capabilities, digital payments giant PayPal has unveiled a new artificial intelligence-powered scam alert system aimed at protecting users from rapidly evolving online fraud.
The rollout, which includes integration with Venmo in the United States, is part of the company’s broader initiative to stay ahead of increasingly sophisticated cybercriminals.
Speaking exclusively to media outlets, PayPal’s Senior Vice President of Global Fraud Prevention, Yigit Yildirim, said the initiative was designed to counter the misuse of generative AI by cybercriminals through more advanced AI tools used for good.
“We need to be smarter and faster than the scammers,” said Yildirim. “We’re combating AI used by bad actors with more intelligent AI used for good.”
The new system uses dynamic machine learning models capable of processing billions of data points. These models can identify unusual behavioural patterns and risk signals even in scams not previously encountered by the system, enabling it to flag suspicious transactions in real time.
Intelligent alerts based on risk level
Unlike traditional fraud alerts that offer generic warnings, the updated system tailors its responses based on the severity of the risk detected. For low-risk indicators, users may simply receive a gentle nudge to double-check the legitimacy of the payment. In more serious cases, where the system has high confidence that a transaction is fraudulent, it can intervene by blocking the payment entirely.
The feature is being implemented globally for PayPal's "Friends and Family" payments and will be made available on Venmo within the US. It is designed not just to detect fraud but to do so without disrupting legitimate user experiences.
Yildirim cited an example to explain the logic behind differentiated alerts.
“If you send the same friend money for lunch every week but always receive the same generic warning, it interferes with your experience and trains you to ignore the warnings,” he said.
The new system was crafted to avoid such ‘alert fatigue’ by offering dynamic responses that adapt to context and transaction history.
Adapting to emerging threats
The technological backbone of this initiative lies in PayPal’s use of adaptive AI, capable of learning from both old and new scam formats. The company says the models can draw connections between known and unfamiliar fraud patterns, thereby flagging risks even in the absence of an exact match.
“With continually learning models, our AI can understand similarities between a known and a new scam,” said Yildirim. “Even if we have not seen that exact same scam before, the AI helps us alert customers to a potential threat.”
PayPal’s decision comes at a time when the cybersecurity landscape is rapidly changing, with the use of AI agents by cybercriminals to launch more precise and damaging attacks.
While the company did not share exact data on fraud rates, the announcement signals a growing trend among major financial technology players to leverage AI not just for convenience but for customer protection amid rising digital threats.
