Human Resources departments across South Africa and abroad are encountering a new development in employee relations: workplace complaints composed with the assistance of artificial intelligence. Platforms such as OpenAI’s ChatGPT enable employees to generate highly structured, legally styled grievance letters in a matter of minutes. These documents often read as though prepared by a seasoned attorney, complete with statutory references, allegations of rights violations, and formal demands. Yet beneath the polished language, the underlying facts may be thin, misunderstood, or inaccurately framed. For HR professionals and in-house counsel, this creates a new and nuanced challenge: separating persuasive drafting from substantive merit.

The rise of AI assisted complaints

Consider a familiar scenario. An HR practitioner receives a lengthy grievance containing dense legal terminology and references to multiple statutes. The tone is assertive and technical, markedly different from the employee’s previous communications. When invited to elaborate in a meeting, the employee struggles to explain the legal arguments set out in the letter and prefers to communicate strictly in writing.

This is no longer theoretical. Employment lawyers in the United Kingdom have reported a sharp increase in AI assisted grievances during 2024 and 2025. Some complaints have even contained fabricated case citations, a known limitation of generative AI tools, which can produce convincing but incorrect information. South African practitioners are beginning to observe similar patterns.

Employees are turning to AI for several reasons. It transforms emotional or fragmented narratives into structured arguments. It gives employees confidence when addressing senior management. It introduces legal terminology that may create the impression of seriousness or urgency.

However, AI systems do not verify facts. They generate probable language based on patterns. A minor interpersonal conflict can be reframed by an algorithm as systemic harassment or constructive dismissal, even where the employee has not consciously adopted that characterisation. The result is often a document that sounds authoritative but may overstate or misstate the true situation.

The legal dimension: Good faith and protected disclosures

When a grievance alleges misconduct such as corruption, discrimination, or unlawful conduct, it may potentially fall within the ambit of South Africa’s Protected Disclosures Act 2000. The Act safeguards employees who report wrongdoing in good faith and who genuinely believe the information disclosed to be true.

An important question arises in the AI era. Can a complaint generated with algorithmic assistance satisfy the requirement of good faith. Artificial intelligence itself holds no beliefs. The critical inquiry therefore remains the employee’s subjective state of mind. If an AI drafted letter accuses an employer of fraud or criminal misconduct, but the employee does not truly believe those assertions, the disclosure may not qualify for protection.

South African law does not shield malicious or knowingly false allegations. Where exaggerated or inaccurate statements are made recklessly, disciplinary consequences may follow. Conversely, the use of AI as a drafting aid does not automatically disqualify an employee from protection. If the employee genuinely endorses the contents of the grievance and believes them to be true, the statutory safeguards remain available. The emphasis is on authenticity rather than authorship.

Practical guidance for HR professionals

AI involvement should not distract from the core responsibility of addressing workplace concerns fairly and objectively. The following steps may assist organisations in navigating AI influenced grievances.

Apply ordinary grievance procedures. Do not dismiss a complaint simply because it appears machine assisted. Follow established internal processes. The factual issue may be legitimate even if the language is inflated.

Engage directly with the employee. Schedule a meeting and ask the employee to explain the events in their own words. This clarifies misunderstandings and confirms whether they stand by the allegations as framed.

Distil facts from rhetoric. Separate emotional or legal embellishments from the underlying incident. If the grievance alleges contraventions of the Employment Equity Act 55 of 1998, identify the specific conduct said to constitute discrimination. Anchor the investigation in evidence.

Seek legal input where appropriate. Where serious legal claims are made, obtain advice from labour law specialists. However, avoid assuming that every AI generated grievance signals impending litigation. The tone may be more dramatic than the reality.

Update internal policies. Organisations may consider including guidance within grievance procedures stating that employees remain responsible for the accuracy of submissions, regardless of drafting assistance. A declaration confirming the truthfulness of allegations can reinforce accountability.

Build AI literacy within HR. Train HR teams and line managers to recognise the strengths and limitations of generative AI. The objective is not to police technology usage, but to ensure informed and proportionate responses.

Preserve the human element. Workplace dispute resolution ultimately depends on communication, empathy, and reasoned dialogue. Technology may assist with articulation, but it cannot replace emotional intelligence or constructive engagement.

A broader shift in employee relations

AI drafted grievances form part of a wider transformation in workplace communication. Employees now have immediate access to tools that elevate tone, structure, and legal vocabulary. This may lead to an increase in formal written complaints simply because producing them has become effortless.

There is also a constructive dimension. If AI empowers employees to voice concerns they would otherwise suppress, organisations gain valuable insight into workplace dynamics. Early awareness of dissatisfaction is preferable to unresolved tension.

The challenge for employers lies in refining investigative skills, assessing credibility, verifying facts, and contextualising allegations without being swayed by stylistic sophistication.

A measured approach in the South African context

In South African labour law, substance prevails over form. A well crafted document cannot transform a weak claim into a strong one. Nor does plain language diminish a legitimate grievance.

The prudent course is neither resistance nor blind acceptance, but informed adaptability. Employers should encourage transparent, informal dialogue before disputes escalate. They should educate employees on responsible AI use. They should reinforce that honesty and good faith remain non negotiable.

Artificial intelligence may assist employees in finding their voice. It does not determine the truth of what is said. At its core, the grievance process remains a human endeavour grounded in fairness, accountability, and genuine engagement.