A responsible boundary between AI feedback and human assessment
AI can help people practise, reflect and prepare, but it should not be mistaken for a full assessment of performance, potential or judgement. The responsible boundary is clear: use AI for structured feedback and human expertise for high-stakes decisions, interpretation and care.
AI is increasingly used to rehearse interviews, review writing, generate prompts and simulate workplace conversations. That makes it useful, but not equivalent to assessment. Stanford’s AI Index shows adoption rising quickly while responsible AI practices still lag behind capability growth 1. In parallel, a Harvard study suggests AI-follower simulations can reveal some leadership behaviours that also appear with human teams 2. The practical lesson is not to replace judgement with automation. It is to use AI as a bounded tool for practice and reflection, while keeping human assessment for context, nuance and consequences.
What AI can responsibly do
AI is most credible when it is used for low-stakes preparation: drafting questions, surfacing patterns in text, rehearsing conversations and giving immediate, repeatable feedback. MIT Sloan’s guidance on generative AI distinguishes between broad productivity tools and tailored solutions, and in both cases stresses the need for user judgement, clear guidelines and critical evaluation of outputs 3. That matters because AI can be fluent without being grounded, and helpful without being complete.
In coaching and development, this means AI can support a person’s thinking before a human conversation, not stand in for it. It can help someone notice habits in their language, test alternative phrasings or structure a reflection. It is less suitable for weighing motivation, values, relationship dynamics or organisational politics, all of which are central to meaningful assessment.
Where the boundary should sit
The boundary becomes especially important when the output may influence opportunity, reputation or wellbeing. Harvard Business School’s experimental findings are useful here because they suggest that performance with AI agents can mirror some aspects of human leadership in a constrained task, but only within a narrow laboratory setting 2. That is evidence of signal, not proof of generalisable competence across jobs, cultures or pressures.
A sensible rule is that AI may inform development, while humans decide on assessment. AI can help a manager prepare for a difficult discussion; it should not determine whether that manager is ready for promotion. AI can highlight communication patterns; it should not be treated as a definitive read on character, judgement or leadership potential. The more consequential the decision, the more important human review becomes.
Why human assessment remains necessary
Human assessment does more than score performance. It interprets context, checks for fairness, notices contradiction and understands what cannot be captured in text alone. Stanford’s reporting on responsible AI is a reminder that capability and governance do not advance in lockstep, and that incidents continue to surface even as tools become more capable 1. For organisations, that means using AI outputs as prompts for inquiry, not as conclusions.
A human assessor can ask what happened before the sample was produced, what constraints shaped it, and whether the person was supported fairly. They can also notice the emotional and relational dimensions that AI cannot reliably infer. In coaching, that matters because growth is rarely just about the visible output; it is also about confidence, self-regulation and the quality of the working relationship.
Practical safeguards for organisations
Set a clear use policy that distinguishes between practice, screening and formal assessment. If AI is used in development, tell people what it is doing, what data it uses and what it does not know. Keep sensitive material out of open tools unless your privacy and governance settings genuinely support the use case. MIT Sloan’s discussion of generative AI governance is helpful here: sanctioned tools, clear guidance and training reduce avoidable risk 3.
Use AI feedback as one input among several. Pair it with manager observation, peer input, work samples and, where appropriate, structured human interviews. Review outputs for bias, overconfidence and missing context. If a tool repeatedly produces neat but shallow feedback, treat that as a signal to recalibrate the prompt or the process, not to trust it more. The boundary is maintained through disciplined use, not through enthusiasm.
A coaching perspective on responsible use
An understated Authenticate advisory perspective is that AI is at its best when it helps a person arrive better prepared for a human conversation, not when it tries to substitute for one. That framing preserves dignity, accountability and trust.
For coaches and leaders, the simplest test is this: if the AI output would be hard to explain to the person receiving it, or hard to defend in a consequential decision, it probably belongs in preparation rather than assessment. Used well, AI can sharpen reflection. Used carelessly, it can flatten judgement into apparent certainty.
Sources
- The 2026 AI Index ReportStanford University — Human-Centered AI · University research · link checked 15 October 2026
- Think You’re a Good Leader? Try Leading an AI TeamHarvard Business School AI Institute · University research · link checked 15 October 2026
- How to manage two types of generative artificial intelligenceMIT Sloan School of Management · University research · link checked 15 October 2026
- Who should be the head of generative AI — and what they should doMIT Sloan School of Management · University research · link checked 15 October 2026
Further professional reading
External professional resources, separate from the numbered sources above. These are not university studies, and listing them does not indicate any collaboration, certification, affiliation or endorsement.
- Global Institute of Directors (opens in a new window)External professional resource · governance and directorship
- World Coaching Institute (opens in a new window)External professional resource · coaching practice