A look at Apate, a conversational bot developed by Macquarie University's Cyber Security Hub in Australia to keep scam callers on the line while collecting data
New scambaiting AI technology Apate aims to keep scammers on the line while collecting data that could help disrupt their business model
Context & Ripple Effects
Apate applies conversational AI defensively: rather than merely blocking suspected fraud, Macquarie University’s Cyber Security Hub uses it to prolong scam calls and capture information that may help investigators disrupt operations. That approach sits against the earlier emergence of generative-AI chatbots tailored to phishing and malware, which lowered the barrier to automating criminal outreach.
The concept also foreshadows Virgin Media O2’s Daisy scambaiter, showing how conversational systems can be deployed as active countermeasures rather than only as filters or warnings.
First-order effects
- Apate can consume scammers’ time while generating call data for the Cyber Security Hub’s analysis, shifting a scam interaction from a one-sided target encounter into a potential intelligence source.
- Scam callers that reach the system face a less efficient calling workflow, while legitimate consumers are kept out of that particular engagement.
Second-order effects
- The model gives telecom operators and anti-fraud teams a template for using conversational agents alongside blocking and detection systems; the later Daisy deployment by a UK mobile operator illustrates that operational direction.
- As scam outreach is increasingly automated, defensive systems will need to distinguish useful intelligence collection from interactions that create privacy, safety, or escalation risks.
Third-order effects
- If scambaiting agents become widely deployed, anti-fraud operations could shift toward automated adversary engagement, with call-level intelligence becoming a more important input to disruption efforts.
- This is part of an AI security arms race: the same conversational capabilities that support criminal phishing and malware tools can be repurposed for defensive deception, making governance and reliable controls more consequential.
The trend: Conversational AI is moving from passive detection toward autonomous, adversarial engagement in fraud defense.