DeepMind team creates human-only application form amid doubts about AI screening

Members of Google DeepMind’s AGI Safety and Alignment Team have raised concerns that the firm’s AI-based applicant screening may miss or delay qualified candidates. To address this, they circulated a form that guarantees a human will read each submission, noting that reviewers are weary of uniform LLM responses. The internal memo, marked with a caution not to share it widely, acknowledges a non-trivial chance of erroneous filtering.

A DeepMind spokesperson affirmed the company’s aim to attract the most qualified applicants and clarified that the form bypasses recruiter review without creating shortcuts to employment. This internal step occurs as AI tools for hiring become commonplace, with companies like Google’s Workspace team promoting similar solutions and competitors such as Workday facing discrimination lawsuits over their AI systems.