Loading story…
Researchers from Amrita Vishwa Vidyapeetham, Ca' Foscari University of Venice, the University of Melbourne, and Ben Gurion University of the Negev found that generative AI chatbots can outperform human scammers during the trust-building stage of pig-butchering fraud. In a week-long simulation with 22 participants, 46 percent complied with an AI chatbot's request to download an app, compared with 18 percent who complied with a human scammer's request to download a game, and participants also assigned higher trust scores to the AI. The researchers said large language models could automate the earliest and longest phase of romance-investment scams, with a human operator stepping in only at the final stage to direct victims to fraudulent investment platforms and avoid model safeguards.
The study was informed by interviews with 145 former scam workers, including trafficking survivors from scam compounds in Cambodia, Myanmar, and Laos, who described criminal use of AI for translation, language polishing, persona development, and video deepfakes. Researchers also found that some leading LLMs could be prompted to impersonate humans and at times conceal that they were AI, raising concerns that scam networks could scale operations while becoming less dependent on visible scam compounds in Southeast Asia and harder for defenders and victims to detect.
Mallory correlates global threat intelligence with your attack surface — know if you’re exposed before adversaries strike.
3 events from the most recent confirmed update back to the earliest known activity.
Based on the experiment and related analysis, the researchers concluded that large language models could automate the early and longest phase of romance-investment or pig-butchering scams. They assessed that a human operator may only need to take over at the final stage to direct victims to fake investment platforms and bypass model safeguards.
The study was informed by interviews with 145 former scam workers, including trafficking survivors from scam compounds in Cambodia, Myanmar, and Laos. Interviewees described scam operations using AI for translation, language polishing, persona building, and video deepfakes.
Researchers from Amrita Vishwa Vidyapeetham, Ca' Foscari University of Venice, the University of Melbourne, and Ben Gurion University of the Negev conducted a week-long simulation with 22 subjects to test whether generative AI chatbots could autonomously handle the trust-building phase of pig-butchering scams. In the experiment, the AI chatbot outperformed human scammers on trust and compliance measures, persuading more participants to download an app or engage with a game.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
3 references tracked. Mallory keeps watching after this page renders.
scworld.com
Open sourcearstechnica.com
Open sourcewired.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
**Transnational criminal networks** are increasingly industrializing online fraud with **AI-enabled social engineering**, according to reporting on scam compounds in Southeast Asia, an Interpol assessment, and policy commentary tied to a new US executive order. Fraud operations linked to *pig-butchering* and romance scams are using generative AI to improve language quality, deepfakes to impersonate trusted people, and low-cost "deepfake-as-a-service" offerings to scale deception. Interpol said AI-assisted fraud is **4.5 times more profitable** than non-AI schemes, while broader reporting describes these operations as structured, multinational enterprises that function like businesses and increasingly rely on automation, synthetic identities, and persuasive impersonation at scale. Reporting from Cambodia and the wider region shows scam operators are now recruiting "**AI face models**" to appear on high-volume deepfake video calls, including applicants from multiple countries seeking work in compounds associated with trafficking-linked fraud operations. The same ecosystem has been described as part of a broader organized-crime model involving forced labor, cryptocurrency investment scams, romance fraud, and impersonation schemes targeting victims globally. One reference on calculating AI ROI in enterprise cybersecurity is **not about this fraud campaign ecosystem**, and an EU sanctions announcement concerns separate state-linked cyber incidents rather than financially motivated AI-enabled fraud.
Jun 29, 2026
OpenAI said it disrupted and banned multiple ChatGPT account clusters tied to online fraud operations likely originating in Cambodia, with some related activity also linked to Myanmar and Nigeria. The campaigns used AI to support **investment scams, romance scams, gambling scams, law-enforcement impersonation, and fake job or task scams**, primarily by generating multilingual messages, fake personas, promotional copy, website and social media content, and internal administrative material. In one newly disclosed case, OpenAI said a coordinated scam network likely operating in or around **Poipet, Banteay Meanchey province** may have interacted with hundreds of targets and included signs of human trafficking and forced criminality, such as recruitment ads for "chatter" jobs and records referencing debts, fines, visas, detention, and escape attempts. OpenAI said the operators relied on familiar fraud playbooks rather than novel AI-enabled techniques. Victims were commonly approached through SMS, job ads, or social platforms and then moved to **WhatsApp**, **Telegram**, or other messaging apps, where scammers used recruiter, mentor, or investment personas to build trust before demanding deposits, cryptocurrency purchases, activation fees, or other payments to unlock supposed earnings or rewards. OpenAI linked some campaigns to repeated employer-style branding such as **Hyesung Advertising** and **Lightning Shared Scooter Co**, said Meta provided at least one tip that helped identify a fake-review job scam, and reported sharing indicators with partners and authorities after enforcement actions.
Aug 5, 2026
A surge in **romance scams** is leveraging **AI-enabled impersonation** to make fraud harder to detect, combining manufactured intimacy with financial theft. Australian police warned more than 5,000 people they may have been targeted in a large-scale operation linked to overseas syndicates, where scammers used mainstream dating apps to initiate relationships and then steered victims into purchasing **fake cryptocurrency**. The playbook described includes rapidly escalating emotional commitment, isolating targets, and pushing conversations off-platform to apps like *WhatsApp* or *Telegram*, reducing victims’ access to in-app safety controls and reporting mechanisms. The fraud techniques increasingly rely on **deepfakes** and automated “AI personas,” undermining traditional verification methods such as requesting custom photos or relying on video calls as proof of identity. Reported tactics include real-time face-swapping and AI voice synthesis during video calls, long-running bot-driven conversations that build trust over months, and “celebrity” impersonation to intensify emotional leverage and extract larger payments. Despite the technology shift, the core mechanism remains psychological manipulation—using scripted narratives and social engineering to move victims from online rapport to off-platform communication and ultimately to financial transfers.
Jun 29, 2026