The technology was enabled by regulatory changes made late last year, which modified how telcos could legally scan the contents of private text messages. “We were always able to monitor metadata, but [by working with government] we were able to get some clarification on what constituted content and what didn’t,” Mr Penn said.

Loading

“Which enabled us to improve the sophistication of our surveillance on SMSs across the network, which means that we can therefore increase our level of blocking.”

Telstra said human specialists would look at blocked texts only very rarely, when the system wasn’t sure if a message was a scam or legitimate, and in those cases the contents would be anonymised.

Adrian Covich, senior director at cybersecurity firm Proofpoint, said automated systems had a role to play in reducing the impact of SMS scams. But more important was tackling the human behaviour that allowed them to spread.

“There’s been a lot of innovation around mobile malware. This isn’t just a guy sitting down, typing each one of these and sending it out,” he said.

Loading

“Particularly we’ve seen a malware variant called FluBot, which has been infecting Android phones, and basically it’s both propagated and funded by these messages.”

FluBot and similar malicious software gets into phones by tricking people into clicking on links, for example by embedding them in messages claiming to be from a bank or courier. Once installed they steal valuable data, but they also use the victim’s phone number to send more messages to others.

Mr Penn said Telstra customers would be able to report suspect messages that made it through, and also had the option to turn off filtering entirely.

Get news and reviews on technology, gadgets and gaming in our Technology newsletter every Friday. Sign up here.