Professor Tatiana Bachkirova of Oxford Brookes University responds to this question: Why is human connection essential in coaching? with this, “Coaching is, at its core, a human-centered process. It depends on skills like empathy, deep understanding, and the ability to interpret and respond to complex, nuanced contexts. These are qualities that AI simply cannot replicate. While AI excels at solving problems based on patterns, in the context of coaching it lacks true understanding, intelligence, or insight. It can simulate conversation but cannot engage in the collaborative and thoughtful processes that define meaningful coaching relationships.” She goes on to say, “If we don’t make human values and connection a priority, coaching could lose what makes it special and meaningful.”
Considerations if using AI:
Establish boundaries to keep trust intact. Boundaries and ethics protect you from accidental overreach. Here are practical boundaries most coaches can follow:
- Do not promise confidentiality if you use third-party tools – this can mean clearly disclosing what you are using to the client to gain their permission first because the Code of Ethics requires confidentiality. Be clear about limits.
- Do not diagnose, give legal or medical advice – and remember, AI does.
- Do not use AI to decide what a client “really means.” Ask the client.
- Ask the client (not AI) for clarification (“What do you mean by that?” and/or “What else do you want to add?” or if you summarize what a client says, “What do you want to correct or change in that summary?”), because overconfidence in interpretation misleads.
- If AI is used with permission, remove identifiable client information and identifying details whenever you can.
- Request options, not one “best” answer, so the client retains control.
A simple prompt template for AI that you can reuse (and edit per client) to prepare looks like this:
- Role (Who are you helping me act as?)
- Context (Client goal and constraints, with no identifiers.)
- Constraints (No diagnosis, no therapy language, and keep it practical.)
- Tone (Warm, direct, short sentences.)
- Task (Give options, then list risks and what to ask the client next)
Structure ideally makes AI useful without handing it the steering wheel.
Privacy, Consent, and Bias, the Non-negotiables
Consent should be plain and in advance. Tell clients when you use AI, what you share, where AI holds the data, and how you store outputs. If you only use AI after sessions for summaries or reflection prompts, say that. If you use it in-session, say that too.
When possible, anonymize details (names, company, location, rare job titles). Use privacy-focused settings or enterprise plans if you have access. Also, delete logs when the tool allows it, and keep your own records consistent with your coaching agreement.
Bias is another risk. AI can slip into stereotypes or one-size-fits-all plans. It may assume a client has the same freedom, money, safety, or culture as the “average” user.
Use a quick stop-and-check test before sharing AI output:
- Does this assume facts not in evidence?
- Can this feel fair across culture, gender, age, and power?
- Does it push certainty where a coach should ask a question?
If any answer is “yes,” revise or discard it.
The 5 principles: transparency, agency, accuracy, care, and accountability.
- Transparency: Inform clients if, when, and how AI is used, and what AI is doing specifically, and get their permission first.
- Agency: The client determines whether AI supports their process.
- Accuracy: The coach verifies AI outputs, personalizes them, and removes wrong assumptions.
- Care: Ensure you protect dignity, avoid harm, and slow down when stakes are high.
- Accountability: The coach provides the process, and the client owns the outcome, not the tool.
Here’s how these principles change a real decision: a coach feels tempted to ask AI to “rate the client’s motivation” from session notes. Under this code, they stop. Instead, they bring it to the next session: “An observation: you made plans twice and didn’t follow through. What got in the way?” That choice protects the client from being reduced to a score, and it keeps the work honest.
Key consideration: AI is creating a risk of cognitive atrophy and a decline in critical thinking if used as a crutch rather than a tool. While AI can lead to a “dumbing down” effect through over-reliance, it does not inherently make humans less intelligent if users maintain active engagement and skepticism of AI-generated output. Considerations:
- Cognitive Atrophy & Dependency: Studies show that reliance on AI for writing and problem-solving diminishes the ability to think critically, with research suggesting that heavy, uncritical use may reduce brain activity and memory recall.
- Reduced “Cognitive Friction”: By eliminating the effort needed for tasks (such as writing or summarizing), AI allows users to skip necessary learning, which can hinder the development of fundamental skills.
- Shifting Definitions of Intelligence: The danger lies in redefining human intelligence to match what AI can produce, resulting in a loss of human-centric creativity and problem-solving.
- AI as an Amplifier: If used correctly, AI can augment human intelligence, handling rote, repetitive tasks and allowing humans to focus on higher-order, creative, and critical thinking.
AI works best when it supports the parts of coaching where a heartbeat is not required. It may help you research, plan questions, and sharpen follow-through. Keep the human parts: presence, safety, and curiosity, in your hands.
If you are going to use AI before, during, or after a session, write a comprehensive disclosure you can provide clients to sign.
When coaches stay responsible for the relationship, trust rises, and results tend to follow.
