The next time you receive a message from an unknown number, pause. The person on the other end might not be a person at all — and they might be better at earning your trust than any human fraudster ever was. A new study suggests that generative AI chatbots are not just assisting scammers; they may be outperforming them at the very core of the crime: building trust.
AI's New Role in the Multi-Billion-Dollar Scam Economy
Fraud operations worldwide steal tens of billions of dollars every year. For years, these operations relied on human callers and message senders to groom victims. Now, researchers say AI is changing that equation.
The study, conducted by academics from Amrita Vishwa Vidyapeetham in India, Foscari University of Venice, the University of Melbourne, and Ben Gurion University of the Negev, examined how generative AI chatbots perform in the scam industry. Their findings point to a troubling reality: AI can sharpen deceptions, polish language, and lubricate banter with victims — often more effectively than a human operator.
Why AI's Ability to Build Trust Is a Game-Changer
Trust is the currency of fraud. A scammer must convince a victim to believe in a fake investment, a false emergency, or a fabricated identity. Historically, this required patience, social skill, and emotional intelligence.
AI chatbots, however, can simulate empathy, maintain consistent stories, and adapt to a victim's responses without fatigue or error. The study suggests that AI may be able to carry out the majority of a long con autonomously — from the first "hello" to the final fake investment pitch.
How the Study Was Conducted
Researchers from the four universities carried out a broad analysis of generative AI's potential in fraud scenarios. While the full methodology remains under review, the core experiment involved testing AI chatbots in simulated scam interactions.
The results indicated that AI could not only replicate human deception but, in some cases, exceed it in building rapport and maintaining engagement over extended conversations. This is a significant departure from earlier assumptions that AI lacked the nuance needed for long-term manipulation.
Who Is Most at Risk From AI-Driven Scams
Anyone with a phone or an internet connection is a potential target. However, the study's implications are particularly severe for vulnerable populations — the elderly, those unfamiliar with technology, and individuals in financial distress.
AI scammers can operate at scale, contacting thousands of potential victims simultaneously. Unlike human scammers, they do not need breaks, do not get tired, and can refine their approach with every interaction. This makes the threat both broader and more personal.
What Cybersecurity Experts Are Saying
While the study's authors have not yet issued public statements beyond the research findings, the implications are clear to fraud prevention specialists. The ability of AI to build trust autonomously represents a new frontier in cybercrime.
Experts in the field have long warned that generative AI would eventually be weaponized. This study provides some of the first academic evidence that the weapon is already operational.
Why This Study Matters Beyond the Headlines
The research is not just about technology; it is about the future of human vulnerability. If AI can outperform humans in building trust, then the traditional advice to "spot the signs of a scam" becomes less effective.
Scam warning signs often rely on detecting human error — typos, inconsistent stories, or emotional manipulation. AI eliminates many of these tells, creating a more polished, more convincing fraudster that never sleeps.
Confirmed Findings vs. What Remains Unclear
Confirmed: The study found that generative AI chatbots can build trust with victims more effectively than human scammers in experimental settings.
Unclear: The full scope of real-world deployment of such AI systems by criminal networks remains unknown. The study is academic and experimental; it does not confirm that large-scale AI-run scams are already widespread.
Speculation: Some analysts believe it is only a matter of time before AI-driven scams become the norm, but this is an inference, not a proven fact.
The Technology Behind the Threat
Generative AI models, like those powering modern chatbots, are trained on vast datasets of human conversation. This allows them to mimic tone, respond to emotional cues, and maintain coherent narratives.
For scammers, this means they no longer need skilled human operators. A single criminal can deploy hundreds of AI agents, each tailored to a different victim profile, all working simultaneously.
Risks and the Need for a Balanced View
It is important to note that AI is also being used to fight fraud. Banks and cybersecurity firms are deploying AI to detect suspicious patterns and block scams in real time.
However, the study highlights an asymmetry: the same technology that protects consumers can be turned against them. The race between AI-powered fraud and AI-powered defense is now underway, and the outcome is far from certain.
A Wider Pattern: AI in Social Engineering
This study is part of a broader trend of AI being used in social engineering. From deepfake voice calls to AI-generated phishing emails, criminals are increasingly leveraging generative tools.
The common thread is trust. Whether it is a voice that sounds like a family member or a chatbot that feels like a friend, AI is eroding the very cues humans use to detect deception.
What You Can Do to Protect Yourself
Given this new threat, traditional vigilance is still your best defense. Be skeptical of unsolicited investment opportunities, especially those that promise high returns with low risk.
Verify identities through independent channels. If someone contacts you asking for money or personal information, hang up and call the official number of the institution they claim to represent. Never rely on the contact details provided by the caller.
What Happens Next in the Fight Against AI Fraud
Regulators and law enforcement agencies are beginning to take notice. However, policy responses often lag behind technological change.
The study's authors hope their research will spur further investigation into AI's role in fraud and lead to better detection tools. For now, the burden of protection remains largely on individuals and the platforms that host these interactions.
Our Take
This study is a wake-up call. For years, we have assumed that scams require a human touch — that a machine could not replicate the subtle art of manipulation. This research suggests otherwise.
The fight against fraud is entering a new phase where the enemy may not be human at all. The most important takeaway is not fear, but awareness. Understanding that AI can build trust is the first step in learning not to give it.
Frequently Asked Questions
Can AI really replace human scammers?
According to the study, AI chatbots can perform the majority of a long-con fraud, including building trust, more effectively than humans in experimental settings. This suggests AI can replace human scammers in many scenarios.
How do AI scammers build trust with victims?
AI chatbots use natural language processing to simulate empathy, maintain consistent stories, and adapt to a victim's emotional responses. They can engage in prolonged conversations without fatigue, making them highly effective at building rapport.
Are AI-driven scams already happening in the real world?
The study is academic and experimental. While there is evidence of AI being used in phishing and other scams, the widespread deployment of fully autonomous AI scammers is not yet confirmed.
What should I do if I suspect I am talking to an AI scammer?
End the conversation immediately. Do not share personal information or money. Verify the identity of the person or institution through official channels, and report the interaction to local authorities or cybercrime units.
