Researchers Use Conversational AI to Combat Rising Phone Scams

An Australian research team has developed a voicebot to engage scammers in long, futile conversations

Researchers Use Conversational AI to Combat Rising Phone Scams
Speech AnalyticsLatest News

Published: September 29, 2023

Charlie Mitchell

A team from the Macquarie University Cyber Security Hub in Sydney has developed a voicebot to combat the rise of phone scams.

Named Apate, after the Greek goddess of deceit, the bot aims to waste scammers’ time by engaging them in long, looping, and ultimately futile conversations.

Researchers applied machine learning (ML) and natural language processing (NLP) techniques to train the voicebot.

In doing so, they analyzed, extracted, and determined script patterns from over 100 scam calls.

From there, the team developed a conversation pattern to drag out those conversations for as long as possible – engaging scammers in numerous rounds of interaction.

It then trained the voicebot on this information and applied voice cloning technology so the bot could adopt a particular persona, emotion, and language.

As such, the scammer makes fewer engagements as they’re kept on the hook by seemingly promising but ultimately pointless bot conversations.

Dali Kaafar, a Professor and Executive Director at Macquarie University Cyber Security Hub, hopes the innovation will bust the business models of the scammers.

“As long as the scammer is convinced Apate is a potential victim, Apate will keep the conversation going,” said Kaafar on LinkedIn.

Should a large proportion of scam calls be directed to Apate, much of a scammer’s time would be spent with the bots without profit, invalidating the scam business model and crippling the scammer ecosystem.

The innovation is somewhat similar to a contact center voice biometrics system, which – when it detects a fraudster – passes them through to a queue.

There, the fraudster sits on hold – not knowing that another agent will not pick up – until they eventually lose patience and hang up.

Another similar anti-scam use case contact centers employ begins with a conversational intelligence system. With this, service leaders dig into scam calls, uncover patterns, and warn agents in real-time.

The trouble is that these systems come at a high price point, which many SMBs can’t afford.

However, suppose telecom companies deploy a bot like Apate across their networks. In that case, there is an opportunity to get ahead of scammers without contact centers having to do their own research.

The telecoms provider may also keep businesses up to date on the latest scams to add a second layer of security.

Meanwhile, the innovation may support everyday consumers too. As Kaafar concluded: “People will not have to suffer the daily grievance of deciphering whether a call is a scam or not.”

Of course, it’s also satisfying to know that the scammer is wasting their time talking to a robot.

The Troubling Rise of Phone Scams

Like many other countries, Australia has reported a surge of scam calls in recent years.

Indeed, Australian authorities endured an astonishing A$3.1BN loss due to scams last year – with phone scams costing Australians A$141MN alone in that period.

The alarming trend has gained momentum due to the proliferation of VoIP technology.

That tech makes it easy to spoof a phone number. Meanwhile, it allows scammers to mask their identity and location.

Moreover, scammers have learned to increase their chances of success by seeming hostile and aggressive. Doing so induces fear from the person on the other end of the line, impairing their rational decision-making.

For instance, a scammer may threaten a consumer by telling them their account is about to be suspended and warning them to act quickly.

Alternatively, a contact center fraudster may appear in a state of panic, telling the agent something may happen to them if they don’t rush to help.

The increasing use of such emotional exploitation has aided the success of phone scams.

Yet, that is something conversational AI may overcome with its emotional independence.

Again, this exemplifies how Apate offers another weapon in the toolkit for telecom companies in combatting fraud – an area they’re currently funneling significant funds into.

For instance, Reuters reports that Telstra – Austrailia’s largest telco provider – spends A$1.2 million a month on scam detection.

With innovations like Apate, telcos will hopefully make more of that colossal budget and better protect people and businesses alike.

For more on the latest virtual agent innovations, read our article: 7 Generative AI Innovations Changing Conversational AI



Conversational AITechnology Media and TelecomsVirtual Agent

Share This Post