New Delhi: Artificial intelligence is making organised online scams faster, more persuasive and increasingly difficult to detect, according to a new international study led by researchers at Amrita Vishwa Vidyapeetham and global collaborators.
The research, conducted in collaboration with Ben-Gurion University, the University of Melbourne, the University of Venice and Colorado State University, draws on interviews with insiders and more than 115 survivors of human trafficking who were forced to work in scam compounds across Southeast Asia.
The study found that many organised scam networks recruit people through fraudulent overseas job advertisements before trafficking them into heavily guarded compounds, where they are coerced into carrying out romance, cryptocurrency and investment scams under constant surveillance and threats.
Researchers said artificial intelligence is increasingly being used to generate personalised messages, improve language, translate conversations and build trust with potential victims, making online fraud more scalable.
One of the studies, titled “Love, Lies, and Language Models: Investigating AI’s Role in Romance-Baiting Scams,” will be presented at USENIX Security ’26, an international cybersecurity conference.
In controlled experiments, participants believed they were interacting with two human operators. However, one of the conversations was handled by an autonomous AI agent. The AI succeeded in persuading participants to comply with requests 46% of the time, compared with 18% for a trained human operator, the researchers said.
The study also found that existing content moderation systems often failed to identify scam conversations because the initial interactions resembled genuine friendship or emotional support rather than fraudulent communication.
Beyond analysing scam operations, the research team has developed a machine-learning model designed to identify fake overseas job advertisements commonly used to recruit trafficking victims. Trained on multilingual datasets in English, Hindi and Bengali, the system detects warning signs such as unrealistic salaries, vague job descriptions and suspicious overseas recruitment offers. Researchers said the technology could eventually be integrated into online job portals and browser extensions to warn job seekers about potentially fraudulent listings.
Dr Krishnashree Achuthan, Dean, PG Programs, Amrita Vishwa Vidyapeetham, said many scam messages are sent by trafficking victims who are forced into online fraud, underscoring the need for stronger awareness and technological safeguards.
The research was released on the UN World Day Against Trafficking in Persons, whose 2026 theme, “Trapped Behind the Scam,” focuses on the role of trafficking networks in cyber-enabled fraud.
