O2 starts blocking suspected scam calls before they reach customers

O2 is rolling out AI-based blocking of suspected fraud calls to all its customers at no extra charge, the company said on Friday. Based on a trial earlier this autumn, it expects more than 3 million malicious calls a week to be stopped in real time.

Until now the network's AI has mainly flagged calls and warned customers about them. O2 says it has flagged more than 1.6 billion suspected scam and spam calls that way since the start of 2024. Under the upgrade, calls suspected of being scams are blocked instead, while those judged to be spam will carry on being flagged.

Customers do not need to do anything. The service is applied automatically on compatible devices, according to ISPreview, and the technology comes from Hiya, O2's existing partner.

How the filter decides

The system watches for patterns across the network, learning from call behaviour and from the reports customers send in. O2 already blocks numbers on Ofcom's Do Not Originate list, and says the AI step widens the range of fraudulent calls it can stop.

The trial also showed how callers respond to being cut off. O2 said 60% of blocked calls were repeat attempts made within a minute of the first, and that a number labelled as fraudulent went on to try three times as many calls, on average, after being blocked as it had before.

What customers can still do

Suspicious calls and texts can be forwarded free to 7726, the reporting number, and those reports feed back into the filter. Text messages are handled by a separate firewall, which O2 says has kept more than 1.4 billion fraudulent messages from reaching customers.

The announcement came with the company's latest quarterly Scam Index. It found that 61% of people who met a suspected scam between July and September had received a suspicious call, five percentage points higher than in the previous quarter.