Google advises job seekers to bypass its own AI hiring tool

Despite marketing artificial intelligence recruitment tools to major corporations worldwide, Google's own researchers have privately advised job applicants to bypass the company's AI-driven hiring system, urging them to complete a special form to ensure their applications reach human recruiters rather than risk erroneous filtering.

According to a confidential document obtained by Bloomberg, Google DeepMind's AGI Safety team acknowledged that the company's AI screening system carries a genuine risk of mistakenly rejecting qualified candidates or delaying their applications from reaching recruiters. To circumvent this risk, applicants have been encouraged to fill out a specific form that routes their submissions directly to human staff, though the document notably cautioned against widely publicising the form's existence. A Google spokesperson, however, maintained that the AI system does not erroneously disqualify candidates, framing the form instead as a mechanism for directing applications straight to particular teams, bypassing recruiters' initial screening stage.

The disclosure comes amid a broader surge in the use of artificial intelligence across corporate recruitment, a trend in which Google's own Workspace division is an active commercial participant, even as the fairness and neutrality of AI-driven hiring practices remain the subject of considerable controversy. A separate Bloomberg investigation reportedly uncovered evidence of name-based discrimination in ChatGPT's outputs, while Workday Incorporated has faced litigation alleging bias on the basis of race, age and disability, allegations the company has firmly denied.

Read More: North Korea deploys AI in cyberattacks: Report

Compounding the issue, job seekers themselves are increasingly turning to AI to draft applications and cover letters, prompting DeepMind researchers to caution in the same form that real human reviewers process submissions and have grown weary of formulaic, AI-generated responses. Applicants were consequently advised to prioritise original, personally written submissions grounded in their own qualifications and experience over AI-assisted drafting, underscoring a growing tension between the technology's promise of efficiency and its practical shortcomings in high-stakes human decisions.