Artificial intelligence (AI) is quickly becoming embedded in everyday workplace processes, from handling HR queries to payroll support. Its ability to provide fast, clear answers makes it an attractive solution for employers looking to improve efficiency and reduce administrative costs and workloads. However, as organisations increasingly rely on AI to support decision-making, it is vital to recognise a key risk: while AI can provide a quick answer, it can still be confidently wrong.

A recent example highlights how even simple queries can lead to incorrect conclusions. When asked about UK VAT deadlines, some AI tools have incorrectly suggested that deadlines automatically move to the next working day if they fall on a weekend or bank holiday. This “sounds right” and aligns with practices in some other countries, but it is not the case in the UK. VAT deadlines remain fixed in law, and submissions or payments must be made on or before the due date, even if it falls on a non-working day. 

This example illustrates a broader issue. AI generates responses based on patterns in data rather than authoritative legal rules. It may draw on multiple sources, some of which are themselves inaccurate, and present a consistent but incorrect answer. When different websites repeat the same misconception, it can reinforce false confidence, making it harder for users to spot errors.

For employers using AI for HR or payroll queries, the implications are clear. AI chatbots can efficiently answer employee questions about payslips, deductions, or leave entitlements. Payroll teams can use AI to explain calculations or flag anomalies. HR systems can provide instant responses on policies or procedures. However, if these responses are taken at face value without verification, there is a risk of spreading incorrect information internally.

The danger is amplified because AI responses are typically well-written and authoritative in tone. When an answer appears clear and confident, users may not think to question its accuracy. In areas such as payroll compliance or employment law, even small errors can lead to financial penalties, employee dissatisfaction, or reputational damage.

Practical tips when using open AI search tools:

• Don’t rely on the first AI-generated answer - cross-check with trusted sources
• Always click through to the original guidance (e.g. GOV.UK or HMRC)
• Prioritise authoritative and professional sources over blogs or forums
• Be cautious of repeated claims - consistency does not equal correctness
• Question answers that sound overly confident or simplified
• Use precise, UK-specific search terms to avoid generalised responses
• Check that information is current and relevant
• Treat AI as a starting point, not a final decision-maker

Staff should also be trained to apply critical thinking when using AI tools and to recognise that responsibility for decisions remains with the organisation, not the technology. 

AI offers significant opportunities to enhance efficiency and service delivery across HR and payroll functions. Used wisely, it can reduce workload and improve responsiveness. But without proper safeguards, it can also introduce risk. The key message for employers is simple: AI is a powerful assistant, but its outputs must always be checked before they are relied on.

This article reflects the position at the date of publication shown above. If you are reading this at a later date you are advised to check that that position has not changed in the time since.   

We regularly publish articles on a range of tax and wider topical issues which affect employers. If you wish to subscribe to our monthly Employer Focus e-newsletter, please contact us.