The workplace challenge: Preventing AI conflict before it starts

Levels of workplace conflict in local authorities are above the national average, says the Advisory, Conciliation and Arbitration Service (Acas), contributing to a wider picture of surging numbers of employment tribunal cases. Lawyers warn that new unfair dismissal claimants could be faced with a five-year wait in seeing a resolution.There are some familiar factors involved in why employees in the sector report higher rates of bullying, harassment and generally unacceptable behaviours, linked back to the tensions caused by the combination of budget cuts and increased demand for services. This is in addition to the particular stresses that come with relationships between the elected councillors on one side and senior officers on the other.We have entered a new world of conflict, where cases are being driven by two developments: a change in the culture of attitudes to making claims – people are more sensitive to what constitutes unacceptable behaviour and more willing to go public with what they have experienced – and, crucially, the support of generative AI as their expert adviser and advocate.Legal firms are reporting that more employees are using AI to check whether there are grounds to make a complaint, exploring every possibility and any similar cases cited on the internet.This results in very lengthy claim documents with little thought or consideration needed on the part of the claimant. Every point needs to be addressed by an employer's response and within tribunal proceedings, meaning a more drawn out time-consuming process for everyone involved.Worse, AI is not accountable for its mistakes and ‘hallucinations'. Software like ChatGPT is designed to maximise engagement by being supportive of the user and their thinking. So, AI is known to exaggerate and stretch material, unconcerned by the lack of facts to back up claims. It may mix up UK and US legal terms and even invent laws.Another concern for local government employers is confidentiality, because any sensitive material entered into open AI tools will from then on become part of the pool of data publicly available to other users.There are basics management and HR need to put in place in response to this new situation to minimise risks: clear guidance on the use of generative AI as part of grievance handling policies, the serious risks involved and serious implications of errors, as well as for GDPR and confidentiality.An added emphasis must be on professionalism as a way of countering AI-assisted claims with consistent management practices in dealing with complaints and making sure there is the appropriate keeping of records, relevant emails, conversations and meeting notes.The challenge from AI is a reminder of the value of informal human conversations as a means of defusing disputes, dealing early with simple misunderstandings, differences of opinion and clashes in terms of ways of working or personality, before relationships are affected. In other words, relying on openness and trust, not artificial systems and processes.That means creating a positive culture of openness and maturity, where employees at all levels are able to generate confidence and trust in each other, based on immediate experience, rather than what is expected or imposed on them.A ‘clear air culture' happens when employees have enough confidence in their authority and each other to speak up and know that they will be listened to, as well as understood in the right way.This kind of workplace culture only comes about with the right skills among staff, particularly line managers and clear, but informal, channels for conversations.Providing access to workplace mediation is one important outlet for ensuring issues get aired, along with making mediation a familiar, more everyday option rather than a last resort.There needs to be attention to levels of psychological safety within teams under pressure, using survey tools to anonymously gather perspectives and experiences, followed up with open discussion of behaviours that may have caused problems in the past and how open conversations and taking measured risks can build trust, along with ideas for building a stronger sense of understanding and safety.In general, there should be a habit of open conversations, both scheduled and unscheduled and the qualities needed to make them ‘good' conversations: listening skills, empathy, curiosity and self-awareness – the self-made trust that makes the difference between knee-jerk irritation and assumptions and constructive, grown-up ways forward.Phil Floyd is head of investigations at workplace conflict specialist CMP. (www.cmpsolutions.com)

Title

AI

The workplace challenge: Preventing AI conflict before it starts

By Phil Floyd | 07 August 2026

SUBSCRIBE TO CONTINUE READING