74% to 51%: What AI Hiring Is Really Costing You

Offer acceptance rates have dropped from 74% to just 51% in two years — and three independent studies from 2025 say AI-driven hiring is a primary cause. Greenhouse found that 70% of hiring managers trust AI to make better decisions, but only 8% of job seekers call AI hiring fair. Gartner puts the candidate trust number at 26%. That gap is showing up directly in your pipeline results. Here's the counterintuitive part: simply disclosing AI use can make things worse. Gartner found that 25% of candidates trust employers less when AI is disclosed without meaningful context. A generic disclaimer isn't transparency — it reads as a warning. What candidates actually want to know is whether a human being is accountable for the decision. An LSE study published in Frontiers in Psychology adds another layer: AI chatbot assessments showed lower predictive validity than traditional psychometric tests. So you may be deploying tools that candidates distrust and that aren't even better at identifying the right people. The fix, according to all three studies, is human-in-the-loop architecture — AI that assists recruiters rather than replacing their judgment. OVI's audio screening chats work exactly this way: structured AI-conducted audio conversations, reviewed by a human recruiter before any hiring decision is made. No video, no biometrics, no facial analysis. Plans start at $99/month, and the approach aligns with NYC Local Law 144, the EU AI Act, and SOC 2 standards.