AI systems provide lower salary targets for women, study finds

A study conducted by researchers from Harvard and MIT reveals that artificial intelligence systems provide gender-biased career guidance. When prompted with identical professional credentials, AI models systematically recommended lower salary baselines for women while suggesting higher compensation targets for men.

In one evaluation, Anthropic’s Claude showed the most significant disparity, proposing nearly $16,000 more for male profiles. OpenAI’s GPT suggested a premium of approximately $2,700, while Google’s Gemini and Meta’s Llama showed smaller gaps. Beyond direct compensation, the models tended to nudge women toward flexible scheduling and caregiving perks rather than financial advancement.

Additional research highlights a broader gender gap in AI engagement. Data indicates that men use AI more frequently in professional and personal contexts and report higher levels of optimism and confidence in the technology. Conversely, women report lower trust in AI, citing concerns regarding sexual deepfakes and job displacement in sectors like administration and customer service. Researchers suggest that the lack of female representation in AI development and leadership may contribute to these biases and the dominance of male-centric needs in AI-generated content.