ChatGPT and other AI tools show age bias in hiring, research finds
ChatGPT systematically excludes older workers from job recommendations, according to research that tested the chatbot's responses to recruitment scenarios. When asked what age group to hire for roles requiring "enthusiasm and new ideas," the tool recommended workers aged 21 to 45 - and did not mention anyone older.
The finding raises urgent questions for HR professionals already using AI in recruitment, performance management, and training decisions. As AI tools become standard in hiring workflows, age discrimination embedded in these systems could affect millions of workers.
How the bias appears in practice
In one test, ChatGPT was asked to draft a job advertisement for an enthusiastic, forward-thinking candidate in the tech industry. The resulting ad used phrases like "fail fast," "fast-moving," and "fresh perspective" - language that signals to older applicants they should not apply.
A follow-up test months later produced a different age cutoff (30s to 50s), showing the bias varies over time. But in both cases, a clear ceiling existed for who could contribute to fast-paced work.
This pattern reflects broader findings. A 2025 survey by the Australian HR Institute found that 18 per cent of HR professionals said they would not hire anyone aged 65 or over "at all." Another 24 per cent considered workers aged 51 to 55 as "older."
Why this matters now
An estimated 58 per cent of workers worldwide use AI tools regularly at work, mostly free generative AI platforms like ChatGPT. As these tools move deeper into hiring systems, the bias becomes harder to detect and easier to scale.
Age bias in AI reflects the training data, design choices, and how organizations deploy these systems. Unlike human recruiters, AI tools can apply biased criteria consistently across thousands of applications.
The legal gap
Australian discrimination laws rely heavily on individuals filing complaints. Fewer than 0.09 per cent of potential age discrimination cases are actually pursued, making it difficult to catch and challenge AI-driven bias.
The European Union has adopted an AI Act that assigns responsibility to developers, providers, and users of AI systems. Most workplace AI applications are classified as "high risk" under that framework.
Australia's current discrimination laws do not explicitly address how organizations must handle bias in automated systems. Positive equality duties - which place responsibility on employers rather than workers to prevent discrimination - exist in Victoria, the ACT, and the Northern Territory, but not consistently across all states or all protected grounds.
What HR teams should do
Automation bias - our tendency to trust automated outputs even when wrong - affects how hiring teams use AI. When filtering job applicants, active oversight is essential.
HR professionals should test AI tools for age bias before deployment, document how systems are used, and maintain human review of all hiring decisions. Discrimination law still applies when using AI, regardless of how neutral the tool appears.
For HR leaders, understanding these risks is part of managing compliance and building fair hiring practices. The AI Learning Path for CHROs covers workforce analytics and recruitment automation strategies that account for these challenges.
Stronger regulation and updated discrimination laws will help, but organizations cannot wait for legal changes. Testing, monitoring, and human judgment remain the most reliable safeguards against age discrimination in AI systems.
Your membership also unlocks:
