A hiring team at a U.S. company using the myHR Partner platform discovered that a reference check it was conducting was not being answered by a real former manager but by an automated system. The team noted that the responses were delayed, uniformly polished, and lacked the natural hesitation that typically accompanies a human conversation. "There was no context beyond the question. Just clean, controlled answers," said a member of the hiring team, according to a report from The Morning Call.

The incident prompted a deeper investigation. The team found that the reference was not a person at all. Instead, the answers were generated by a large‑language‑model‑based tool that produces reference letters and other HR documents. Similar AI reference letter generators are available from several vendors, including Genbaz, Typli, and Smodin. These tools use generative AI to produce polished, professional‑sounding reference letters based on user‑provided details.

The use of AI in reference checks is part of a broader trend of automation in recruitment. AI can screen resumes, schedule interviews, and even predict candidate success. However, the technology also introduces new risks. Reference check fraud—where candidates provide false or fabricated references—has long been a concern. Recent research on fraud detection highlights several indicators: duplicate IP addresses, unusual device activity, and unverifiable referee identities.

"Detecting reference check fraud requires a multi‑layered verification framework," said a researcher in the field. "AI can help flag inconsistencies, but it also can be used to create convincing fake references, as this incident shows." The hiring team’s discovery underscores the need for robust verification processes. Some vendors now offer reference fraud checks that track IP addresses and device signatures to confirm that a response originates from a legitimate source.

The incident also intersects with regulatory developments. The European Union’s Artificial Intelligence Act, which entered into force on 1 August 2024, imposes obligations on AI systems used in professional contexts, including hiring. High‑risk AI applications—such as those that influence employment decisions—must meet transparency, security, and quality requirements and undergo conformity assessments. Limited‑risk applications are required to provide transparency notices. The Act also creates a European Artificial Intelligence Board to oversee compliance.

U.S. companies that use AI for recruitment, even if they are not subject to the EU Act, face similar scrutiny from industry bodies and privacy regulators. The General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) both require that personal data used in AI systems be handled with care, and that individuals be informed about automated decision‑making.

The myHR Partner platform, part of Scout Talent’s suite of recruitment tools, offers background checks, reference verification, and compliance support. Scout Talent’s website lists several partners that provide AI‑enhanced background checks and fraud detection services. "Our goal is to give recruiters confidence in the data they rely on," said a Scout Talent spokesperson.

As generative AI tools become more sophisticated, the line between human and machine responses in reference checks will blur further. Recruiters are advised to combine AI‑generated content with human oversight and to employ technical safeguards that can detect synthetic or fabricated responses. The industry is also exploring the use of synthetic data to train AI models while preserving privacy.

In the coming months, several AI vendors are expected to roll out new features that flag potential AI‑generated references. Meanwhile, regulatory bodies may issue guidance on best practices for using AI in hiring. Until then, recruiters must remain vigilant and verify references through multiple channels to mitigate the risk of fraud.