Applying for a job at Google? DeepMind staff warns its AI hiring filters may screen out your CV ‘incorrectly’

Alphabet Inc.'s Google markets its artificial intelligence tools to businesses as a means to expedite recruitment by aiding employers in sifting through vast quantities of job applications and pinpointing promising candidates. Nevertheless, certain researchers at Google DeepMind's AI division are adopting safeguards against the utilization of such technology in their own hiring procedures.

The AGI Safety and Alignment Team, which focuses on mitigating risks linked to advanced AI, has instructed applicants for their open positions to fill out an additional form alongside their job applications. This measure aims to prevent candidates from being disqualified by Google's internal AI-driven screening systems prior to their applications being reviewed by a human recruiter, as per a document obtained by Bloomberg.

The document contained a cautionary note, "Please do not share this doc widely." The form informs applicants that Google's application system has a "non-trivial probability" of erroneously rejecting their CVs or causing delays in transmitting them to the team. "Our application system has a non-trivial probability your CV will be screened out incorrectly or take too long to reach us," the document asserts.

"Filling out this form ensures a real human on the team will see your application." A spokesperson for Google DeepMind declared that the company's aim is to "recruit and hire the most qualified talent at Google DeepMind" and dismissed the notion that its systems inaccurately filter candidates. "This team set up a separate form to bypass the recruiter review and have their resumes directed to the team members.

But there are no shortcuts to getting hired," the spokesperson added. The adoption of AI in recruitment has surged significantly, with some human resources departments apprehensive that automation might eventually diminish the necessity for certain hiring positions. However, the manner in which companies employ these systems remains inconsistent and often elusive to job seekers.

Some enterprises harness AI models to rate candidates, whereas others leverage automated systems to scan CVs for specific keywords and filter applicants accordingly. Google's Workspace division, which provides products such as Google Drive for businesses, advertises its most recent AI functionalities as resources capable of assisting HR departments "in saving HR time by rapidly generating job posting drafts, evaluating resumes, and forecasting hiring requirements."

AI-driven hiring systems have also faced criticism for potentially discriminating against candidates. Bloomberg's investigation revealed that OpenAI's ChatGPT displayed potential biases based on applicants' names. Meanwhile, Workday Inc., a seller of workplace management software, is currently entangled in a lawsuit alleging that its AI hiring systems screen applicants based on race, age, and disability in contravention of the law.

Workday has denied the allegations, stating that humans make the hiring decisions. The company did not respond to a follow-up inquiry for further comment. Simultaneously, certain job applicants are leveraging AI to their advantage by manipulating the filters or inundating the system with applications at an unprecedented pace. The Google DeepMind team remains vigilant about this possibility and advises candidates that their applications would be more compelling without the assistance of AI.

The form designed to bypass automated screening includes a warning: "A real human will read these. These humans get really tired of reading LLM answers, because they all sound very samey."