Deepfake scam networks are now hiring human faces on telegram

There is a job posting circulating on Telegram right now that asks applicants to submit daily photos, conduct up to 150 video calls per day, and disclose their vaccination status and marital situation. No company name. No contact details. Just a set of software tools, a strict schedule, and the implicit understanding that you will be spending your working hours pretending to be romantically interested in strangers in order to steal their money. This is not a fringe phenomenon. It is an organised labour market for deepfake-assisted fraud, and it is growing faster than most people realise.

Telegram as a recruitment engine for scam operations

Wired's investigation into dozens of Telegram channels has exposed what amounts to a functioning hiring pipeline for criminal tech outfits. The channels publish structured job listings — formatted like legitimate remote work offers — targeting individuals in Turkey, Russia, Ukraine, Belarus and several Asian countries. Researchers at ChongLuaDao, a Vietnamese non-profit focused on scam prevention, have confirmed that these listings circulate openly, with minimal obstruction from the platforms hosting them.

The mechanics are straightforward. Recruits are provided with AI-powered face-swapping software. They sit in front of a camera, their expressions and movements captured in real time, and the system maps their gestures onto a synthetic persona. That persona then conducts romantic scam operations — building emotional relationships with victims before extracting money or personal data that can be leveraged in secondary fraud schemes.

What these listings actually demand from applicants

The working conditions described in these postings would be alarming in any legitimate industry. Four and a half days off per month. Mandatory daily photo submissions. Video and audio message creation on demand. One listing specifies approximately 100 video calls per day; another pushes that figure to 150. These are not casual gigs. They are high-throughput fraud operations with the operational rhythm of a call centre.

Authenticity is the product being sold, which is why the listings are oddly specific about presentation. Wigs are banned outright. Filters are permitted, but only if the result looks convincingly real. Live-action video is preferred. The recruits are not just faces — they are performance assets, and the criminal networks running these operations understand that a convincing human presence dramatically increases the success rate of romantic scams.

The deliberate ambiguity that makes prosecution difficult

Here is where it gets structurally clever. The majority of these postings never explicitly mention fraud. They describe the role in vague terms — AI model work, video interaction roles, content creation — and bury the red flags in plain sight. Multilingual applicants are preferred. Cryptocurrency experience is valued. The absence of any identifiable employer is treated as normal.

A subset of listings drops the pretence entirely and openly solicits workers for «love scam» operations, requesting candidates with years of experience on crypto fraud platforms. But even these are rarely actioned by platform moderators, which tells you everything about the enforcement gap that these networks are exploiting.

Cybercrime researcher Hieu Minh Ngo of ChongLuaDao has noted that the listings move through Telegram's ecosystem with virtually no friction. The infrastructure is decentralised enough to be resilient, and the recruitment model is scalable. You do not need a sophisticated operation to run this. You need software, a camera, and a workforce willing to look into a lens for ten hours a day.

The human cost sitting behind the technology story

It is tempting to frame this purely as a deepfake story — another chapter in the AI misuse narrative. But the more uncomfortable truth is that this is also a story about economic desperation being industrialised. The people responding to these listings are not, in most cases, sophisticated criminals. They are individuals in economically precarious regions being handed a toolkit and pointed at victims. The criminal networks absorb almost none of the legal risk. The recruits absorb most of it.

Meanwhile, the victims at the other end of those 100 daily video calls are real people being manipulated by what they believe is a genuine human connection. The deepfake layer adds a dimension of technological sophistication, but the underlying mechanism — exploiting loneliness and trust — is as old as fraud itself. What has changed is the throughput. One operator, with the right software and a willing face, can now run dozens of simultaneous relationships across multiple time zones. That is not an evolution of romance scamming. It is an industrialisation of it.