The organisations seeing the greatest benefit are often not those with the most advanced technology, but those with clear governance, effective training, strong accountability and a commitment to human oversight and judgement.
AI is already embedded in everyday workplace decisions
AI is becoming a routine part of working life. Government research found that one in three people had used AI at work in the previous month, despite most workers receiving no AI-related training. As adoption accelerates, organisations are increasingly using AI to improve efficiency, reduce administrative burdens and support activities such as drafting communications, analysing data and streamlining recruitment. However, while AI can process information quickly, it cannot exercise judgement, understand context or assess fairness in the way humans can.
Recruitment remains the first major risk area for employers
The use of AI in recruitment is now widespread, from sourcing and screening candidates to analysing interviews and ranking applicants. While these tools can improve speed and consistency, employers must ensure that recruitment decisions remain fair, transparent and explainable.
The key risk is not just bias, but the inability to explain how a decision was reached. AI tools may also create liability under the Equality Act 2010 where they disadvantage individuals sharing a protected characteristic. In practice, this is most likely to arise as indirect discrimination, where an apparently neutral criterion places a particular group at a disadvantage.
Regulators are also paying increasing attention to the use of automation in recruitment. The Information Commissioner's Office (ICO) has identified automated decision-making in recruitment as a key area of focus, particularly in relation to transparency, discrimination and the responsible use of personal data – please see ICO update.
Employers should therefore ensure there is meaningful human oversight of recruitment decisions and that hiring managers can explain not only what decision was made, but why.
AI enhanced job applications
AI tools can help candidates produce polished CVs, tailored applications and sophisticated interview responses. While there is nothing inherently wrong with this, employers may find it harder to distinguish between an individual's actual skills and their ability to use AI effectively. In some cases, capability gaps may only become apparent after employment has commenced, increasing the importance of robust assessments and effective probation management. The risks may increase further from January 2027, when the qualifying period for unfair dismissal is due to reduce to six months and the cap on compensation will be removed.
AI should inform decisions, not determine them
While recruitment dominates much of the discussion around AI, the greater employee relations risk often arises once someone is already employed.
Managers are increasingly using AI to summarise information, draft documents and support decisions relating to performance, capability, attendance and conduct. Problems arise when managers treat AI outputs as answers rather than inputs. A performance summary generated in seconds may appear persuasive, but managers still need to question whether it is accurate, balanced and supported by evidence.
Employees are more likely to challenge a decision that affects them than the use of AI itself. Where disputes arise, employers should be able to show that AI informed the decision, but that the final outcome was reached through meaningful human review and judgment.
AI is changing the nature of grievances
Employees can now use publicly available AI tools to produce lengthy, highly structured complaints in a matter of minutes. These grievances often look sophisticated and legally detailed. However, volume does not necessarily equal substance.
The practical challenge for employers is that AI-generated grievances can make it harder to identify the core issues requiring investigation. Complaints may contain irrelevant points, inaccurate references or legal arguments that appear convincing but are unsupported. Employers should nevertheless approach these complaints with care, as grievances often provide the first indication of concerns that later develop into discrimination, whistleblowing or constructive dismissal claims.
Employers should not dismiss grievances simply because AI may have been involved. Instead, they should focus on identifying the issues genuinely in dispute, keeping investigations proportionate and ensuring compliance with the Acas Code on Disciplinaries and Grievances and internal procedures. Informal discussions can often help identify the real concern more quickly than lengthy written submissions.
In some cases, it may be helpful to ask the employee to prepare an executive summary of a lengthy grievance at the outset of the process, highlighting the key allegations, issues and outcomes sought. This can help maintain focus and ensure the investigation remains proportionate.
Employers should also consider whether workplace mediation may offer a more effective route to resolution, particularly where the grievance stems from a breakdown in working relationships rather than a dispute about facts or misconduct.
AI is beginning to reshape workplace disputes
AI is also affecting litigation.
As with grievances, employees can now use readily available AI tools to quickly produce lengthy tribunal claims. Some employers are reporting claims that are broader, more legally complex and more difficult to narrow at an early stage than would previously have been the case. While these documents can appear well-researched and persuasive, they may also contain inaccuracies, unsupported allegations or legal references that do not withstand scrutiny.
As claim time limits increase from three to six months from October 2026, and the qualifying period for unfair dismissal reduces, employers may see more claims being brought and more issues being raised. Combined with existing tribunal backlogs, this risks creating a perfect storm of increased claim volumes, longer-running disputes and higher defence costs. Employers may increasingly consider challenging claims or allegations that are unsupported or have no reasonable prospects of success – whether by seeking further and better particulars or making an application for strike out.
AI at work: Why culture and governance, not technology, will determine success
The greatest AI risks rarely arise from the technology itself. More often, they emerge where AI is adopted without clear accountability, oversight or understanding. As AI becomes embedded in recruitment, employee relations and workplace decision-making, employers need to ensure that managers understand how to use these tools appropriately and when human judgment must take precedence. Just as importantly, organisations need to foster a culture in which employees feel confident questioning AI-generated outputs, escalating concerns and exercising independent judgment rather than treating technology as infallible.
The challenge is not whether organisations should use AI, but whether they have the right framework in place to use it responsibly. That may include reviewing AI policies and procedures, training managers, establishing approval processes for new tools and maintaining clear audit trails for AI-assisted decisions. Employers should also consider whether their grievance, investigation and litigation processes are equipped to deal with increasingly sophisticated AI-assisted complaints and claims. A strong workplace culture can act as an important safeguard, helping to ensure that AI is used consistently with organisational values and ethical standards, rather than simply for efficiency or convenience.
Ultimately, successful AI adoption is not about restricting innovation. It is about ensuring that AI supports human decision-making rather than replacing it. Organisations that invest in governance, accountability, training and AI literacy will be best placed to realise the benefits of AI while managing legal, employee relations and reputational risk. Those that combine these measures with a culture of transparency, challenge and responsible decision-making are likely to be the most successful, creating an environment where AI enhances, rather than undermines, trust in the workplace.
