BrightHR flags legal gaps in AI-written HR policies

BrightHR says AI-generated HR documents can look complete while missing the practical detail employers need to apply rules fairly and safely. The company recommends every AI draft be reviewed by a human HR and employment law expert before use.

Why it matters: - AI tools are increasingly being used to draft HR policies, contracts and handbook content. - BrightHR says that can save time, but it can also create legal and operational risk if employers rely on the output without review. - Missing detail can affect consistency, fairness and how safely a policy stands up in practice.

What happened: - BrightHR shared findings from an internal review led by Gemma O’Connor, Head of HR Advisory and Technical Services at BrightHR. - O’Connor assessed an AI-written absence management policy generated from a simple prompt. - The test prompt asked the AI tool to create an absence policy that said employees would receive a formal warning after a third absence. - BrightHR said the policy looked credible at first glance, but lacked key operational detail underneath the headings.

The details: - The AI policy told employees to notify their manager “as soon as possible” if they could not attend work. - BrightHR said that wording is too vague to apply consistently across different start times, shift patterns and critical roles. - O’Connor said policies should set a clear deadline, such as contact by the employee’s start time. - BrightHR said the policy also failed to spell out who should be contacted if a manager is unavailable. - The review found no guidance on whether text messages are acceptable. - The policy also did not say whether third-party notifications, such as a colleague calling in, are allowed. - O’Connor said text messages and third-party notifications should not be acceptable. - The AI policy included a “formal warning threshold” saying a third absence in a rolling 12 months would trigger a formal warning. - BrightHR warned that this kind of wording can lead employers to treat discipline as automatic. - O’Connor said a warning should follow a fair, thorough process, not a literal trigger without notice, investigation or review of circumstances. - BrightHR said AI wrote what it was asked to write, not necessarily what the business needed. - The company said HR policies also need to align with other procedures, organisational consistency, relevant legislation and established workplace practice. - BrightHR said the core issue was not sentence quality, but missing context.

Between the lines: - The review suggests AI can produce polished drafts that still miss the details that matter most in disputes. - That gap matters because vague wording can leave employers exposed when absence management becomes contentious. - The findings also show why trigger-point policies need human judgment, not just automated language. - O’Connor said AI in HR can be positive if businesses use it with caution and the right training.

What's next: - BrightHR recommends any HR document generated by AI, or drafted in any other way, should be reviewed by a human expert in employment law. - The company advises employers to treat AI drafts as a starting point, not a finished policy. - BrightHR also recommends building in timeframes, escalation routes and reporting channels. - The company says employers should embed fair process into trigger-point rules, align AI content with existing documentation, check it against legislation and workplace practice, and train anyone using AI for HR tasks. - For more information, BrightHR provided contact details for Miriam Payne at BrightHR, +447929711809 ext.