Google DeepMind’s AGI Safety team warned potential applicants on 12 August 2026 that the artificial intelligence used for resume pre-selection in its recruitment process may be error-prone, potentially causing qualified candidates to be wrongly rejected by automated filters. To minimize this risk, the team recommended using a separate form to ensure manual review by human staff, bypassing the automated stage so that applicants’ qualifications are assessed directly by people rather than relying on the company’s algorithms. This internal precaution contrasts with the official position of parent company Google, which denied reports of insufficient reliability in its systems and disputed that systematic errors occur in automated filtering, maintaining the effectiveness of its hiring tools. The concerns reflect a broader debate over AI-driven HR processes, which face growing criticism for potential bias and lack of transparency, with parallels drawn to bias issues in models like ChatGPT. The software provider Workday was cited as a prominent example of legal risks, having been sued in the past for alleged discrimination by its algorithms, highlighting the danger that automated systems may disadvantage certain applicant categories due to flawed data patterns.