Are Employees Treating AI as a Superior Adviser?

A manager sends an employee a private email about their performance. The employee is upset and wants a second opinion. They paste the entire email into an AI chatbot, including the manager’s name and details of a client complaint, and ask whether the criticism is fair. Within seconds, the chatbot suggests possible motives, recommends a course of action and drafts a reply.

Are Employees Treating AI as a Superior Adviser? A worker glances back while sharing a confidential email with an AI chatbot.

The employee went looking for a second opinion. Now the chatbot’s reading of the email could shape what they say to their manager and whether the matter escalates. That’s where the question “Are Employees Treating AI as a Superior Adviser?” comes in. The employee has trusted AI with both the judgement call and the private email it was based on. What does that mean for the people who expected the correspondence to stay confidential?

From Impressing the Boss to Asking AI What to Do

In Are Employees Secretly Using AI at Work to Impress the Boss?, we looked at how employees can quietly use AI to write polished emails, prepare better reports and finish work faster. A manager sees the improvement and may assume the employee has developed new skills or found a more efficient way of working.

The same employee might eventually ask AI whether a manager is being unreasonable, whether a colleague intended to undermine them or whether a disagreement warrants a formal complaint. A tool they first used to improve their writing is now helping them interpret other people’s behaviour. If they have had months of good results using AI to draft documents, it is easy to see how that trust might carry over into decisions about people.

What Happened to the Australian Workers?

In its report on Australian workers using AI, Daily Mail Australia described how a South Australian solicitor told a court she had dismissed her paralegal after AI was used to prepare documents for a family law appeal. The filing contained inaccurate or nonexistent legal authorities that went unchecked before submission. The court referred the lawyers involved to their professional regulators.

The Fair Work Commission’s decision in the Fujifilm case describes how a disagreement over a Microsoft Teams message escalated into a workplace dispute that ended in employee Marcus Wibmer’s dismissal. Wibmer used AI to help write emails that Deputy President Tony Slevin described as repetitive and demanding. Slevin said AI appeared to have given Wibmer false confidence that the correspondence was appropriate.

In another case covered by the report, the commission found that a worker had been bullied. It also observed that lengthy, accusatory correspondence that appeared to have been prepared with AI contributed to the breakdown in relations. An employee can have a genuine grievance and still make it harder for someone to understand what happened if the strongest evidence is buried beneath pages of accusations.

What Happens to Confidentiality When Employees Share Emails with AI?

Imagine writing privately to a colleague about a health issue, a complaint or a mistake involving a client. Would you feel comfortable sharing the same details if you knew the recipient might paste your message into a personal AI account and ask the chatbot to analyse what you meant?

A performance review may refer to other team members. A grievance might include details of someone’s health. Even an informal Teams exchange can identify a client or reveal a decision that has yet to be announced. Removing names may still leave enough information in dates, job titles and descriptions of incidents for someone to recognise the people involved.

The Office of the Australian Information Commissioner’s guidance on AI at work advises organisations it regulates to avoid entering personal information, particularly sensitive information, into publicly available AI tools. How a service handles what an employee shares depends on the product, account and privacy settings. A chatbot conversation may feel private to the employee even when their organisation has little control over what happens to the information they entered.

The effect on working relationships can begin as soon as people suspect their messages are being shared. A colleague might think twice about explaining why they are struggling. A manager might stick to formal wording even when a more personal response would help. If people become guarded about what they put in writing, everyday conversations can become harder long before any complaint reaches HR.

Why Employees May Treat AI as a Superior Adviser

AI can give someone space to explain a difficult situation without interruption. It can organise events into a timeline, suggest questions for a meeting and help an employee find the words to raise a sensitive issue. Asking a colleague may feel awkward, particularly when they know the manager involved.

An experienced colleague might ask what happened before the email arrived, whether the same message went to the whole team or how the manager normally communicates. A chatbot may have only the email and the employee’s account, yet it can still produce a convincing explanation of what the sender supposedly intended.

We explored this in Is AI Making Us Feel Smarter Than We Actually Are?, where we looked at how easily a confident answer can be mistaken for understanding. In a workplace dispute, a polished response can make an employee more certain about another person’s motives without giving them any new evidence.

Keeping Judgement and Confidentiality in the Workplace

Someone who wants help preparing for a conversation can often ask a general question without sharing the original correspondence. “What questions should I ask to clarify critical feedback in a meeting?” can produce a useful list. If the advice depends on the exact wording of a confidential email, it may be time to speak with someone authorised to review it in confidence.

Employers also have decisions to make. If they encourage staff to use AI, employees need to know which tools are approved, what information they can share and when a person needs to review the work. They also need a credible route for discussing sensitive issues. If someone can get a patient hearing from a chatbot in seconds but cannot speak privately with anyone at work, it is easy to see where they may turn the next time a difficult email arrives.

Explore our courses for a better understanding of AI in the workplace from all angles, including practical skills, confidentiality, professional judgement and responsible use.

Published 21 November 2026.