Recruiting is facing a new threat: a surge in fraudulent job applications generated by artificial intelligence. In response, leading ATS platform Gem has announced a partnership with Tofu, a startup specializing in identity verification. The alliance embeds an automated tool for flagging suspicious profiles directly into the hiring process. With thousands of client companies, Gem is now giving its users a shield against fake candidates, polyworkers, and bad actors exploiting AI to infiltrate organizations.
The Gem-Tofu Partnership: A Technology Response to Fraud
Tofu, a Toronto-based startup founded two years ago, pivoted last September to AI-powered identity verification. Its technology analyzes candidates' social profiles by cross-referencing several signals: account age, posting activity, number of LinkedIn connections, and consistency of information. A typical fraudulent profile is a four-month-old LinkedIn account with only two or three connections. The Gem integration lets recruiters receive an automated report flagging suspicious profiles from the earliest stages of the process. The solution addresses three types of fraud: mass AI-generated applications, polyworking involving multiple simultaneous full-time jobs, and malicious infiltration aimed at stealing data. Backed by a $5 million funding round, Tofu is extending its reach to thousands of Gem's client companies.
How AI Detection Works
Tofu's technology uses machine learning to analyze candidates' digital behavior. It examines the age of social accounts, interaction frequency, likes, posts, and professional connections. This data is compared against models of authentic profiles to generate a trust score. Recruiters then receive a detailed report flagging anomalies, letting them prioritize manual checks on suspicious applications instead of wasting time reviewing every single file.
The Benefits for Gem Customers
The thousands of companies using Gem can now add Tofu verification as an add-on module to their hiring process. This native integration eliminates technical friction and enables fast adoption. HR teams save valuable time by automatically filtering out fraudulent profiles before interviews. The solution also reduces security risks tied to malicious infiltration, which is especially critical in sensitive sectors like tech and finance.
The Different Forms of Application Fraud
Gem's Steve Bartel and Tofu's Jason Zoltak identify three levels of fraud. The first, relatively harmless, involves desperate job seekers using AI tools to apply to jobs en masse. These automated tools often misrepresent candidates' real experience by generating generic resumes. The second level involves polyworking: people simultaneously holding down multiple full-time jobs, like Indian developer Soham Parekh, discovered in July 2024 working for several Silicon Valley startups at once. The third and most dangerous level involves bad actors. The US Department of Justice has indicted several people for helping North Korean IT workers land remote jobs to fund their country's military program.
Mass AI-Generated Applications
Facing a tough job market, some candidates use automated tools to apply to hundreds of postings a day. These tools tailor resumes to match ATS keywords but often produce inconsistent applications. Recruiters end up with unmanageable volumes of low-relevance applications. This creates a vicious cycle: the more candidates use AI to game the ATS, the more recruiters rely on AI to filter applications, widening the gap between real talent and real opportunities.
Malicious Infiltration
The most serious cases involve identity theft and industrial espionage. Bad actors create fake profiles or steal photo-less LinkedIn identities to apply for jobs. Once hired remotely, they gain access to internal systems to extract sensitive data or trade secrets. Cases involving North Korean IT workers show how this kind of fraud can serve geopolitical goals, turning recruitment into a national security threat vector.
Practical Advice for Recruiters and Candidates
Bartel and Zoltak recommend that professionals keep their LinkedIn profiles up to date with a recent photo, to avoid being flagged as suspicious or becoming victims of identity theft. Recruiters should build identity checks into the earliest stages of the process, especially for sensitive or fully remote roles. Tools like Tofu make it possible to automate this screening without adding to the workload. Companies should also train their HR teams to recognize red flags: recent profiles with no history, inconsistencies between a resume and social profiles, and reluctance to do video calls. Finally, legitimate candidates should understand that these measures protect the integrity of the process and support fair competition in the job market.
Conclusion
The Gem-Tofu partnership marks a decisive step in the fight against AI-fueled application fraud. By combining the power of a leading ATS with innovative identity-verification technology, this alliance gives recruiters the tools they need to navigate a job market transformed by artificial intelligence. Protect your hiring process today by integrating automated fraud-detection solutions.
