Optus is hoping more refined AI and machine learning modelling will help it to detect and isolate scam text messages in near real-time.
Speaking at a recent Optus Tech Day, senior director for data services engineering, Belinda Lofts said the telco aimed to be "relentless in ensuring that we're reducing scam SMS from reaching our customers”.
It will do so in part by boosting investment in its AI and machine learning capabilities, which it views as key tools to combat scams.
“We will be looking at enhancing our machine learning models to include more content [detection] and we will start to explore the introduction of AI as well," Lofts said.
“Partnerships will continue to be key for us because we think that we cannot just do this alone.
"We need to work collectively together to solve for these problems, so they will be key."
Lofts hoped the extra investment would allow Optus to "get ahead" of the scammers, allowing it to block a greater proportion of messages from being delivered to customers' devices.
Optus already has a number of tools in place to protect its customers from receiving voice or SMS scams, according to Lofts, including “an SMS firewall that sits in front of the platform,” which “does a lot of the filtering and blocking of any scam messages for us.”
Firewall rules are based on a blacklist of identified scam numbers as well as on alpha tags or alphanumeric sender IDs.
“This is where the sender ID comes through as a name or a word rather than the mobile number itself," Lofts said.
"We can actually block these from overseas networks hitting customers here in Australia as well.”
Lofts said Optus also employs "smart content filtering analysis that allows us to actually identify the scam messages from the legitimate messages."
Other measures in place include partnering with various Australian governments and looking “for malicious URLs, particular phrases potentially or even callback numbers”.
“I think one of our big observations is that as soon as we implement effective rules to block scam messages, we see the scammer has changed the content of the messages. They're continuing to evolve and adapt," Lofts said.
"We need to be ahead of them.”
An Optus spokesperson told iTnews the company “continues to evolve our capability to bring in better and wider sources of scam intelligence, rolling out smarter automatic detection capabilities, and enhanced information sharing. “
“Automation, AI and machine learning capabilities are all important features of improved scam protections," the spokesperson said.
"The technology will be able to effectively distinguish between legitimate messages and scam messages, in near real-time.
"These capabilities allow quicker scam identification and quicker actions.”
The spokesperson added the speed of identification and blocking “is vital to effective scam protections."
“Scammers are constantly evolving and change their approach as soon as blocks are put in," the spokesperson continued.
"Effective protections require the solutions to be created and implemented as quick as the scammer."
Optus is also partnering with the Australian Financial Crime Exchange (AFCX) on a new feature that the agency is developing called the anti-scam intelligence loop, which is able to identify current scam websites and numbers for the telco to block.
Optus has also developed ‘call stop’ which enables the company to inform users a phone number they are attempting to call is a potential scam.
