AI bias in hiring, LLMs, and employment decisions
Key Questions
What evidence shows AI bias in hiring and employment tools?
2026 benchmarks indicate GPT-4o and Gemini 2.5 Pro exhibit racial and gender bias in outputs, while many employers lack visibility into AI tools embedded in HR systems used for hiring and promotion.
What new regulations address AI in hiring?
Canada's Bill C-36 introduces regulatory requirements for AI use in HR, while the EU AI Act's reactive approach has drawn criticism for insufficient protections against bias in employment decisions.
How do structural factors contribute to AI bias in hiring?
Gender and racial biases in algorithms are reinforced by existing structural inequalities in training data and decision-making processes, perpetuating disparities in hiring and credit scoring.
Multiple reports highlight persistent bias in AI systems used for hiring, promotion, and credit scoring. 2026 benchmarks show GPT-4o and Gemini 2.5 Pro exhibit racial and gender bias. Many employers lack visibility into AI tools embedded in HR systems. Canada's new AI in HR rules (Bill C-36) add regulatory requirements. Gender bias in algorithms is reinforced by structural inequalities, and the EU AI Act's reactive approach is critiqued.