Microsoft's AI Fights Phone Scams, Puts You in Control

Microsoft wants to give its artificial intelligence (AI) the ability to analyse telephone conversations to detect and protect you from telephone scams. Dubbed Azure Operator Call Protection, the service is capable of analysing and scanning all exchanges between you and the other caller.

In a new blog post, Yousef Khalidi, CVP of Azure for Operators, highlighted exciting updates about Microsoft's Azure for Operators suite of cloud-based solutions that enable Microsoft's AI and Copilot solutions users to elevate customer experiences, streamline business operations, monetise 5G investments and modernise their network.

It is worth noting that Azure Operator Call Protection can analyse conversations in real-time and alert the user if the caller is trying to scam them. This AI-powered anti-scam solution was developed in collaboration with BT Group, a London-based telecom provider.

"We've already implemented various measures to protect our customer base from many millions of scam calls and texts, but with fraud a growing problem in the UK, we are always looking at how we can take consumer protection to the next level," Reza Rahnama, MBE, MD Mobile Networks, BT Group said.

"That's why we are collaborating with Microsoft to pilot Azure Operator Call Protection and analyze how, by using advanced (AI) capabilities, we can further enhance our defenses against fraudulent activity," the top executive added.

Microsoft is currently demonstrating how the technology works at the Mobile World Congress conference (MWC) in Barcelona. The company's rival, Google is also beta testing an enhanced version of Pixel Call Screen with conversational AI to detect spam calls.

How AI can combat phone scams

If the AI detects elements suggesting fraud, it will interrupt the call with a warning. Suspicious elements that trigger the AI include specific words used by the caller or requests to share personal information.

Once the warning is triggered, the user will be informed of the system's suspicion through a message. Users can choose to continue the call if they believe it's legitimate, as Microsoft explains that the AI does not autonomously halt conversations. So, the control will remain with the user.

Aside from receiving an alert message, the user will see a summary explaining why the call might be a scam and the AI will even suggest precautionary measures. Microsoft says it neither stores the content of the analysed conversations on its servers nor uses it to train its AI.

Cybercriminals rely heavily on phone calls to trick users into sharing their bank details, credit card numbers and other sensitive information. They then use this data to make online purchases for you or steal your money.

To recap, UK police busted phone scammers and arrested 142 people in 2022, as part of their biggest ever counter-fraud operation. Scammers usually impersonate trusted entities like customer service representatives, government leaders and bank officials to target vulnerable individuals, particularly the elderly.