A small study from Northeastern University points to a troubling possibility: artificial intelligence may strengthen one of the internet's most ruthless scams.
In a seven-day experiment designed to mimic "pig butchering" scams, in which fraudsters build long-term relationships with victims before persuading them to hand over money or access to their finances, an AI chatbot convinced nearly half of participants to make a risky financial move. The chatbot significantly outperformed the human scammers involved in the test.
What's happening?
According to E Gamers, researchers at Northeastern University wanted to see whether people or AI were better at winning strangers' trust online. To do so, the study pitted AI bots against humans in a style of scam often referred to as pig butchering.
Pig butchering scams depend on time and emotional manipulation. A fraudster builds a relationship over days or weeks, then guides the target toward something harmful, such as a fake investment, a malicious app, or another damaging action. The experiment was built around that same pattern.
In the study, 22 participants interacted with either human scammers or an AI chatbot. Each scammer had seven days to build rapport before making the same request: install an app or try an online game. In the study, the app or game stood in for the point at which, in a real scam, someone might be pressured into risking their device or finances.
The human scammers got about 18% of participants to go along. The chatbot persuaded nearly half.
TCD Picks » Quince Spotlight
💡These best-sellers from Quince deliver affordable, sustainable luxury for all
The researchers said this did not require a purpose-built scam system. The model used in the test was Claude, and it was given a simple instruction not to disclose that it was AI.
Co-author Yisroel Mirsky, a computer science professor at Ben Gurion University, said, as reported by E Gamers, "With relatively little effort, we're able to make an agent that can outperform a human at building this exploitable emotional trust,"
Why does it matter?
Suspicion was not enough to keep many participants from going along with the request.
Among the 22 participants, 20 correctly identified the scammer as AI. Even so, many still complied.
TCD Picks » Quince Spotlight
💡These best-sellers from Quince deliver affordable, sustainable luxury for all
That is especially concerning at scale. Even after recognizing the AI, many participants still followed through, and while a person can handle only a limited number of conversations at once, a chatbot can manage dozens or even hundreds without becoming tired, distracted, or inconsistent.
AI's growth also carries consequences beyond scams. The same technology that can help utilities balance electricity demand, improve grid efficiency, and support cleaner energy systems also depends on power-hungry data centers that consume large amounts of electricity and water for cooling.
AI may offer some benefits, but it also places significant strain on the grid, raising security concerns and potentially contributing to higher costs that are passed on to consumers through their energy bills.
What can I do?
Instinct alone is not enough. Even when people recognized the bot, that did not always stop them from complying.
AI-driven scams require the same caution as other forms of online fraud. If someone you meet online asks you to download an app, accept a game invitation, or move money — especially when you have never met them in person — it is safest not to do so.
Get TCD's free newsletters for easy tips, smart advice, and a chance to earn $5,000 toward home upgrades. To see more stories like this one, change your Google preferences here.
