> ## Content Index > Fetch the complete content index at: https://digitalteachernotes.com/llms.txt > Use this file to discover other available public pages before exploring further. # What Happens When AI Starts Rating the Teacher? - URL: https://digitalteachernotes.com/what-happens-when-ai-starts-rating-the-teacher/ - Published: 2026-09-30T03:26:38.000Z - Updated: 2026-09-30T03:26:38.000Z - Description: A UK training company is using AI to review every instructor session. The company says it improves coaching quality. Some teachers say it feels like constant surveillance. - Author: Digital Teacher - Tags: News, AI Teachers are used to being observed. A manager sits in on a lesson. Maybe once a month. Maybe once a term. Then comes feedback. But one UK training company is trying something very different: **AI that reviews every teaching session.** Multiverse, a technology and apprenticeship training company, has introduced an AI system called **Compass** that analyzes instructors’ online lessons and scores aspects of their teaching. The company says the goal is to help coaches improve. Some instructors say it feels more like being watched all day. ## What does the AI actually look for? Compass analyzes transcripts of online teaching sessions against a teaching-quality framework. According to Multiverse, it can identify patterns and highlight sessions that may need human review. Reporting by *The Guardian* says the system can flag things such as: - filler phrases - hesitation - unclear instructions - long pauses - possible digressions - technical issues that take too long to resolve - extended one-to-one exchanges that pull attention away from the group It then produces a status and confidence score for managers to review. [Read Multiverse’s explanation of Compass](https://www.multiverse.io/blog/how-were-using-ai-to-invest-in-our-coaches?ref=digitalteachernotes.com) ## From one observation a month to every session Multiverse says the system solves a real problem. Previously, a manager might observe an instructor roughly once a month. That gives only a small sample of what their teaching is actually like. Compass can review every session instead. From the company’s perspective, that means: - more consistent quality checks - earlier detection of problems - more personalized professional development - less time spent by managers manually collecting evidence Multiverse says the AI does not make employment decisions. Humans still review the information and decide what to do. ## Teachers describe a very different experience Some instructors interviewed by *The Guardian* say the system has created intense anxiety. They describe the monitoring as “remorseless” and “unnerving.” Several said they became more conscious of how their words might appear in a transcript than of responding naturally to students. One concern is that normal teaching behaviour can look strange when reduced to text. A teacher may appear to “digress” because they answered a student’s unexpected question. A pause may be deliberate thinking time. An extended conversation with one student may be exactly what that student needed. The AI sees a pattern. The teacher sees a classroom. Those are not always the same thing. ## This is a very different kind of EdTech Most education AI is aimed at students or teachers as a tool. Generate a quiz. Summarize a text. Create a lesson plan. Explain a concept. Compass represents something else: **AI as evaluator.** That raises a much harder question. What happens when the system is not helping the teacher teach, but judging how well the teacher is teaching? ## Observation changes behaviour There is an old problem with lesson observation. People behave differently when they know they are being watched. Multiverse itself acknowledges that traditional observations can feel artificial because teachers may “perform” for the observer. AI was supposed to solve that by observing normal teaching instead. But continuous monitoring may create the opposite problem. If every session is analyzed, teachers may start adapting to the system. Use fewer filler words. Avoid pauses. Avoid unexpected discussion. Stay closer to the script. That may produce cleaner transcripts. It does not necessarily produce better teaching. ## Can AI judge context? This may be the central problem. Teaching quality is full of context. A good teacher may slow down because a class is confused. They may abandon a planned activity because students need something else. They may let a discussion run because it has become unusually productive. They may repeat themselves. They may pause. They may improvise. A rigid evaluation system can interpret those deviations as mistakes. A human observer can ask: **Why did you do that?** An AI system has to infer the answer. That is a much harder task. ## Multiverse says humans remain in control The company strongly rejects the idea that Compass replaces human judgment. It says AI performs the initial analysis, while managers make decisions. Multiverse also argues that the system allows managers to spend less time gathering evidence and more time helping instructors improve. The company says its learners give coaches a **97% satisfaction rating**, and that Compass is part of an effort to keep teaching quality consistently high. That is an important distinction. The system is not, according to Multiverse, an automated hiring-and-firing machine. It is supposed to be a tool for directing human attention. ## But the psychological effect still matters Even if a human ultimately makes the decision, constant automated evaluation can still change the working environment. If teachers feel that every hesitation, phrase or deviation may appear on a dashboard, that can affect: - confidence - experimentation - spontaneity - willingness to adapt - willingness to take risks in teaching And that leads to a bigger question for education. **Can a system designed to standardize teaching quality accidentally standardize teaching itself?** ## This probably will not be the last example AI is already being used to evaluate: - customer-service calls - sales conversations - employee productivity - interviews - workplace communication There is no reason education will be exempt. Schools and training organizations will increasingly have the ability to analyze huge amounts of classroom data automatically. That means the important question is no longer: **Can AI evaluate teaching?** It clearly can evaluate some measurable features of teaching. The harder question is: **Which parts of good teaching should never be reduced to a score?** ### Sources [The Guardian: Teachers at Euan Blair firm report ‘horrendous stress’ after AI used to rate their work](https://www.theguardian.com/technology/2026/sep/28/teachers-euan-blair-firm-report-horrendous-stress-ai-used-to-rate-their-work?ref=digitalteachernotes.com) [Multiverse: How we’re using AI to invest in our coaches](https://www.multiverse.io/blog/how-were-using-ai-to-invest-in-our-coaches?ref=digitalteachernotes.com)