Knowledge

Employee grievances range in seriousness from minor complaints, such as unwanted feedback around missed deadlines, issues concerning different working styles, or personality clashes, through to the most serious cases of harassment and discrimination.

This briefing does not seek to question the genuineness of an employee's complaint or the impact it has on their work life and wellbeing, nor does it negate the requirement for an employer to conduct a fair and impartial investigation to resolve any issues.

However, we have heard from several HR professionals in the past weeks and months about how the use of AI, by disgruntled employees, to assist with drafting internal complaints, grievances and tribunal claims, has escalated matters that could otherwise have been resolved informally, has hardened positions, and has led to lengthy formal complaints that can go beyond the facts of the dispute.

Without critical thinking, the use of AI to draft complaints can complicate the resolution process and potentially lead to:

• long and complex documents which take hours to unpack and where the real root of the issue may not be clear, as the employee themselves may not understand what the AI has drafted for them;
• legal jargon being included, typically with references to specific provisions of employment law, breach of trust and confidence, or psychological safety (with references often being to UK or US law rather than the law applicable in Jersey or Guernsey);
• validation bias or the ''illusion of truth", where AI hallucination or AI creation of additional grievances (that were not originally in the contemplation of the employee) become the 'truth' as the employee 'hears' the AI repeatedly tell them how they have been wronged and this affirmation then creates the perception of truth;
• positions becoming more entrenched and adversarial at an early stage, and relationships beginning to break down, as there is no informal discussion and so no chance to resolve matters through constructive conversation;  
• the employee seeking unrealistic outcomes, e.g. UK or US based outcomes; and/or
• large volumes of email/written correspondence, as the employee typically refuses to meet in person to discuss the issues and requests all communication to be in writing so they can use AI to analyse it and craft a response.

Employee considerations

In the past, employees may have sought guidance from a trusted confidant or raised matters informally with HR to talk things through. This would typically allow a conversation to take place, where issues could be resolved.  

We have heard from several HR professionals that this step is being skipped, with the employee instead simply using AI tools to generate a formal complaint.  The AI doesn't challenge the employees' assumptions, and most employees don't ask the AI to look at it through the lens of the company and be critical of and challenge their own account.  As a result, matters escalate into a formal dispute, which could otherwise have been resolved informally.

This can have unintended consequences, as the nature of these processes can erode trust and result in a breakdown of the working relationship.  AI-generated documents can also result in significant lost time and productivity, which affects revenue within the business. 

It can also have inadvertent consequences, where AI inflates allegations impacting the reputation of another employee (which in certain circumstances can expose the employee to claims of defamation).  It can also result in potential disciplinary action, where the employee discloses confidential information to open/public AI tools or, alternatively, where an employee uses company AI software during working hours, when they are being paid to do their job.

It is human nature to avoid conflict and difficult conversations, so it can seem like an easier solution to use AI and to hide behind an email rather than have a difficult conversation, but employees should carefully assess what outcome they are seeking and whether that is best achieved through open and constructive dialogue or via formal written communication. They should also review any AI generated content to ensure it accurately and genuinely reflects the facts and their opinion.

What can organisations do about it?

A sound, and well communicated, approach to AI more broadly, can encourage a measured approach by employees towards the use of AI, which will translate into a more grounded use of AI in the case of complaints and disputes. The company should revisit its risk appetite in relation to AI use and consider its preferred approach in line with its culture, risk tolerance and technological capability.

Employers should review all applicable policies and procedures and consider including clauses to:

• prohibit covert recording of meetings (which can be uploaded to AI);
• require employees to disclose when they have used AI;
• require employees to disclose AI prompts used when creating work product produced by AI;
• set parameters on acceptable and unacceptable use of AI, which may include detailing tasks, setting time limits or being required to interrogate/challenge the AI with specific prompts;
• prohibit company confidential information, legal advice, or personal data about employees being uploaded into open-source AI;
• make clear that AI use on company systems is company property and is disclosable (on request and in any potential litigation); and
• require employees to attend an in person (or video) meeting to discuss AI-generated content (typically in the context of grievances or other workplace disputes) where the content that has been created is long, complex, unclear or contains lots of legal references, to allow open discussion in order to clarify the issues and the outcome the employee is looking for.

It should be drawn to employees' attention that failure to adhere to the new policies could, in appropriate circumstances, result in disciplinary action being taken against the employee, in particular where the employee's work product is misleading, or where allegations are not genuinely held (i.e. they are a product of AI), unfounded or vexatious.  

Any proposed amendments to policies and procedures should be viewed through the lens of a person with a protected characteristic, to ensure that there are no potentially unintended adverse consequences. For example, AI can assist people to clarify their thoughts and present a more structured work product, which means that arbitrarily limiting access to AI may place certain persons at a disadvantage.  

Employers may be required to balance the rights of various individuals.  The aggrieved employee will often cite stress as being a factor which has led them to raise their complaint formally, with AI assisting them to clarify and articulate their thoughts. However, often the use of AI and the complaint that is produced, including where the aggrieved employee insists on all communication being in writing (so they can run it all back through AI), results in a significant increase in the volume of work for others within the organisation, such as the HR team, senior managers and the leadership team, and this can then lead to those individuals being overworked or stressed.  Employers have a duty of care to all employees, so may wish to carry out a risk assessment to understand potential effects so they can seek to address them.

Before making any changes to policies, consider if the changes would be non-contractual or a change to the contract of employment, and whether they are a requirement or desirable (from an engagement perspective). It is also advisable to consult with employees, or groups of employees (e.g. a manager forum), in relation to the changes before they are introduced.  

Whether consultation is required or not, a clear, concise and unambiguous communications plan should be prepared so that all employees are aware of the changes/new policies. 

As with all new policies, training should be provided, particularly to managers on how to spot AI (mis)use and, where identified, how to have what can be a difficult face to face conversation.

For financial services employers, it will also be appropriate to look at the guidelines issued by the various regulators (both for the organisation and where individual employees hold relevant professional qualifications for which an additional professional body provides guidance) and it may be appropriate to include additional clauses in your policy and/or consider reporting requirements to applicable regulators.

The Employment law team at Bedell Cristin have experience of working in in-house, onshore and offshore roles, and can advise HR professionals on resolving workplace disputes and on updating policies, procedures and employee handbooks on the use of AI.