How AI is changing the face of workplace communications

Individuals and businesses increasingly turn to AI for advice across many areas of life, and the workplace is no exception. Indeed, as many employers increasingly require employees to use generative AI tools (such as ChatGPT, Google Gemini, and Microsoft Copilot) at work, it is no surprise that many of those employees are using the same tools for their own personal benefit. Employers are seeing a surge in employees using generative AI to draft workplace communications, submit formal grievances, data subject access requests (DSARs), and litigate disputes. This trend, described by some legal professionals as “GIT” (Grievance Invented by Technology), is creating distinct operational, legal, and relationship challenges for HR teams and managers.

THE MAIN CHALLENGES

Inflation and “hyper-legalisation” of complaints
Historically, a grievance might have centred on a disagreement with a manager or other employee, or dissatisfaction with a workplace decision that required a practical resolution. Today, relatively minor or informal workplace friction is increasingly converted into lengthy, formalised documents. When an employee feeds a routine complaint into an AI model, the software often reframes the concern using legal language, for example, alleging “breaches of implied trust and confidence”, “whistleblowing”, “discrimination”, or “victimisation”, even where the facts do not support these legal concepts and where the employee might not actually have characterised their complaint as such.

Inaccurate law and hallucinated precedents
There is increasing evidence that AI models introduce errors, outdated law, or hallucinated facts into grievances, DSARs, and Employment Tribunal claims. As a results, employers can encounter:

• Citations of non-existent English court or tribunal cases
• Misapplied statutory tests
• Americanised terminology and US legal concepts that do not apply in English employment law
• Legal conclusions asserted as fact before being properly established
• Demands for outcomes that are disproportionate to the issues raised and/or beyond what an Employment Tribunal can realistically award.

Entrenched positions and “AI certainty”
Because AI models are designed to be agreeable to user prompts, employees who use AI to “sense-check” an issue often receive output validating their grievance. This creates a false sense of security regarding the strength of their case or potential financial settlement. Consequently, employees become resistant to informal resolution, mediation, or reasonable compromise, adopting adversarial positions early in the process.

Data protection and confidentiality breaches
To generate detailed workplace communications, employees may copy and paste internal company emails, or sensitive personal data about colleagues, or commercially sensitive information into public AI platforms. This exposes employers to UK General Data Protection Regulation (GDPR) and Data Protection Act 2018 compliance risks, data leaks, and potential confidentiality breaches.

Covert AI note-taking and meeting-recording
Employees are increasingly using AI transcription tools (such as Otter.ai or Fireflies) or embedded smartphone AI tools during informal catch-ups, disciplinary hearings, and grievance meetings. This presents challenges around:

• Breaches of internal communication and confidentiality policies
• Unconsented processing of third-party personal data disclosed in confidential hearings
• Disputes over the accuracy of AI-generated meeting summaries presented as official records.

Disproportionate strain on HR resources
Under the current Acas Code of Practice on Disciplinary and Grievance Procedures (Code), employers must investigate workplace complaints thoroughly and fairly, regardless of whether they have been generated by AI. Sorting through lengthy, AI-generated narratives to extract core factual claims requires substantial HR and management hours, often inflating internal administrative burdens and legal expenses.

Increased tribunal filings and procedural delays
AI significantly lowers the technical barrier for unrepresented individuals (including those using new, and sometimes unregulated, AI “law firms”) to draft formal legal filings, such as claims, witness statements, and legal submissions. This ease of drafting has contributed to higher volumes of Employment Tribunal submissions, compounding existing tribunal backlogs in England, which are only likely to increase as key new rights granted under the Employment Rights Act 2025 come into force in the coming months. In some regions, full hearings are currently being listed a year  and, sometimes, significantly more ahead. Employers are also having to spend more time and money requesting case management orders to narrow down or strike out speculative claims.

PRACTICAL STEPS EMPLOYERS CAN TAKE

Update policies and guidelines

Most organisations now have policies covering IT and data protection. These should be updated to address AI use specifically. Employers should:

• Ensure documentation is clear and comprehensive.
• Clearly define what data employees may and may not input into external AI tools, and state that uploading confidential business information, trade secrets, or colleagues’ personal data to public AI tools constitutes a breach of IT policy and UK GDPR/Data Protection Act obligations and, therefore, a disciplinary offence.
• Update handbook policies regarding meeting protocol, explicitly prohibiting unauthorised AI transcription or recording during workplace meetings.
• Consider creating internal online portals (with word limits) through which grievances and DSARs must initially be submitted.

Clarify complaints early

When a multi-page, hyper-legalised document lands on HR’s desk, employers should seek to schedule an early clarification session to verify which claims the employee genuinely understands and personally endorses, helping to separate genuine concerns from AI-generated content. Detailed notes should be kept of any conversations or meetings with employees.

Pinpoint the desired resolution early

AI tools often generate lists of grievances without specifying actionable remedies. Employers should ask the employee directly: “What specific outcome are you seeking from this process?” Identifying target resolutions early focuses the investigation and can keep expectations grounded.

Distil and group allegations

Employers are not required to address every submission in a lengthy grievance and should instead consolidate them into core factual themes (e.g. workload, communication issues, pay dispute) and where possible, seek agreement on that scope before investigating.

Prioritise conversations over correspondence Exchanging endless letters or emails risks initiating an “AI prompt war”, where both sides rely on software to generate defensive correspondence. Employers using AI-assisted drafting should also remember that prompts and outputs may be disclosable in DSARs or litigation. Hearings and interviews should be conducted in person or via live video where possible.

Verify legal precedents early

Employers should not assume cited legislation or case law is correct, and should check with legal counsel before responding.

Adhere to the Code

Regardless of how a grievance is generated, the employer’s statutory obligation remains unchanged: listen, investigate fairly, and reach a reasoned decision. Adhering to the Code ensures that an employer’s process remains legally defensible. Indeed, Acas is currently consulting on the terms of a new Code, which is expected to place greater emphasis on informal resolution – something that may assist employers dealing with GITs.

The use of AI is already a common feature in workplace relations, and this is only likely to increase in the future. Used well, it can help employees to understand their employment rights and help them communicate concerns more clearly. However, used badly, it can unrealistically inflate expectations, escalate disputes, breach data protection rules and, ultimately, result in lengthy and expensive Employment Tribunal litigation.

It is important that employers take steps to ensure that they have effective systems and practices in place for dealing with AI-assisted grievances, DSARs, and claims. Maintaining a positive workplace culture and addressing concerns promptly remain the most effective ways to reduce conflict and support early resolution.

If you have questions on your AI policies or would like assistance managing complex AI-assisted grievances, DSARs or employment claims, please contact Nick Tsatsas.