What happened
Wired reports that Apate, an Australian company, has deployed a system of approximately 350,000 AI bots designed to intercept and engage with phone and chat scammers. These bots, used by banks and supported by telecoms, are trained to maintain conversations for extended periods—often over two hours—without falling for the scam, thereby wasting scammers' time and collecting real-time intelligence on fraud operations.
Apate, named after the Greek goddess of deception, has spent the last two years developing a platform that diverts phone scammers onto calls with AI bots. The company's founder, Dali Kaafar, describes the bots as 'perfect victims' designed to keep scammers engaged while collecting intelligence. The platform is currently used by banks and supported by telecom companies, with a reported fleet of around 350,000 bots.
The bots are engineered to have distinct personalities, language skills, and profiles to avoid detection. They mimic human behavior, such as occasionally hanging up or expressing skepticism, to maintain the illusion of a real conversation. Wired's Kernel Panic newsletter tested a demo version of the system, finding the AI responsive and natural in timing, though the testers were unable to convince the bot to invest in a fake cryptocurrency opportunity within six minutes.
Beyond phone calls, Apate's bots infiltrate online scam chat groups and respond to text messages. The company claims to have collected over 250,000 pieces of real-time information about fraudsters, including scam URLs, money mule accounts, and bank details. This data is intended to aid in the disruption of scam operations and improve intelligence sharing among stakeholders.
The article contextualizes Apate's work within a broader trend of using AI to disrupt cybercriminals. It references research by Mark Vero at ETH Zurich, who found that LLM-powered honeypots keep AI agents attacking systems significantly longer than traditional honeypots. This suggests that AI-driven deception is becoming a more effective tool for defenders against both human and automated threats.
Why it matters
This development marks a shift from passive defense to active disruption in the fight against cybercrime. By leveraging to create convincing, skeptical 'victims,' Apate aims to reduce the volume of successful scams by exhausting the resources of industrial-scale scam operations. The system also generates valuable data, such as scam URLs and money mule accounts, which can be shared with law enforcement and financial institutions. This approach addresses the limitations of traditional law enforcement in tackling cross-border digital crime and offers a scalable, automated method for protecting potential victims at scale.
The deployment of AI bots to waste scammers' time represents a practical application of in cybersecurity. By engaging scammers in prolonged, unproductive conversations, Apate aims to reduce the number of potential victims reached by these criminals. This is particularly relevant given the industrial scale of modern scam operations, which use automated dialing to reach millions of people.
The intelligence gathered by these bots provides a new data source for law enforcement and financial institutions. Real-time information on scam URLs and financial accounts can help in takedown efforts and the identification of money laundering networks. This addresses a key challenge in combating cybercrime: the difficulty of tracking and prosecuting criminals across borders.
The use of AI in this context also raises questions about the ethical and legal implications of interacting with criminals using automated systems. While the goal is defensive, the methods involve deception and the collection of data from criminal actors. The effectiveness of these tools depends on their ability to remain undetected, which may lead to an arms race as scammers develop ways to identify AI bots.
This development highlights the dual-use nature of AI in the security landscape. While AI can be used to enhance the capabilities of attackers, it is also being leveraged to create more robust defenses. The integration of LLMs into honeypots and scambaiting tools suggests a future where AI-driven deception plays a central role in protecting digital infrastructure and individuals from cybercrime.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').Why can ethical evaluation not be reduced to one model score?
What to watch next
Monitor the adoption of AI-driven honeypots and scambaiting tools by major financial institutions and telecom providers. Watch for regulatory responses to the use of AI in interacting with criminals, as well as any evidence of scammers adapting their tactics to detect and bypass these AI systems. Additionally, observe the effectiveness of intelligence sharing between private AI security firms and public law enforcement agencies.
Track the expansion of Apate's platform and its adoption by other major banks and telecom providers. The scale of deployment will determine the overall impact on scam volumes and the quality of intelligence gathered.
Monitor research and industry reports on the effectiveness of AI-driven honeypots and scambaiting tools. Look for independent verification of the claims made by Apate and other companies regarding the duration of engagements and the utility of the collected data.
Watch for regulatory developments regarding the use of AI in interacting with cybercriminals. Governments may need to establish guidelines for the collection and sharing of intelligence gathered from criminal operations using automated systems.
Observe the tactics of scammers to see if they adapt to detect and bypass AI bots. This could involve the development of new techniques to identify non-human interlocutors, which would require continuous updates to the AI systems to maintain their effectiveness.