Updated
Updated · Yahoo Finance · Aug 10
Google DeepMind Team Bypasses HR AI Filters for Job Candidates as AGI Safety Unit Flags Screening Risk
Updated
Updated · Yahoo Finance · Aug 10

Google DeepMind Team Bypasses HR AI Filters for Job Candidates as AGI Safety Unit Flags Screening Risk

3 articles · Updated · Yahoo Finance · Aug 10

Summary

  • Google DeepMind’s AGI Safety and Alignment Team is asking applicants to submit a separate form alongside standard applications so strong candidates are not automatically filtered out by Google’s internal hiring AI.
  • The workaround appeared in a document viewed by Bloomberg that warned, “PLEASE DO NOT SHARE THIS DOC WIDELY,” indicating the team sees enough risk in the screening system to route candidates around it.
  • That stance undercuts Alphabet’s broader pitch to corporate customers that AI can speed up recruiting by sorting large volumes of applications and identifying the most promising prospects.
  • The episode highlights a wider tension inside AI companies: teams focused on advanced-model safety are showing limited trust in the same automation tools their parent companies market for workplace use.

Insights

Why did Google's own AI safety team secretly tell job applicants to bypass the very resume-screening algorithms Google sells to other companies?
If the creators of advanced AI do not trust it to read a resume, how many qualified candidates are algorithms silently rejecting worldwide?