Teachers at Multiverse report 'horrendous stress' under AI monitoring

Instructors at Multiverse, the £1.6bn tech training company founded by Euan Blair, say an AI system that analyzes transcripts of their online classes and flags suspected mistakes to managers has produced "horrendous stress" and a feeling of being constantly watched.

The allegations, first reported by The Guardian in an exclusive investigation published on 28 September, come from staff who deliver apprenticeship training in computer and AI skills to thousands of workers across the UK public and private sectors. According to the teachers who spoke out, the AI system scans transcripts of online classroom sessions, scores their teaching, and alerts human managers when it suspects instructors are doing something wrong.

Teachers at the company described the system as "remorseless" and "unnerving," according to the report. The strongest phrase — "horrendous stress" — came from staff describing the cumulative effect of knowing every session they teach is subject to automated review.

The story matters beyond one company. Multiverse is one of the most prominent players in the UK's apprenticeship training market, and the case is an early, concrete example of AI-driven workplace surveillance being applied to knowledge workers — not warehouse pickers or call-center agents, but professionals whose judgment and interpersonal skill are the core of the job. If AI monitoring spreads from logistics and customer service into teaching, consulting, and other judgment-heavy professions, Multiverse's experience becomes a test case for how far employers can push automated performance evaluation before it degrades the work itself.

What the system does

According to the Guardian's reporting, the mechanism works as follows: online classroom sessions are transcribed, and an AI model analyzes those transcripts. The model assigns scores and is programmed to alert human managers when it detects instructors suspected of "doing things wrong." Human managers then act on those alerts.

That architecture — automated flagging with human follow-up — has become a common pattern in AI workplace tools. The teachers' objections, as reported, center not on the follow-up but on the continuous, always-on nature of the automated layer. Staff told the Guardian they felt "constantly watched," and that the pressure never lifted between sessions.

The instructors who spoke to the Guardian characterized the monitoring with two words that carry weight precisely because they are theirs, not the reporter's: "remorseless" and "unnerving."

The company behind the system

Multiverse was founded by Euan Blair, son of former Prime Minister Tony Blair, and has grown into one of Britain's most valuable education technology companies, with a reported valuation of £1.6bn. The company's core business is apprenticeship training — delivering skills programs in areas including computer and AI skills to workers at thousands of UK public and private sector employers, subsidized through the UK's apprenticeship levy system.

That business model places Multiverse in a regulated training market where quality assurance and compliance matter commercially. An AI system that watches every session, transcribes it, and flags suspected problems is, from an operations standpoint, a cheap and scalable way to enforce consistency across a large roster of instructors teaching at scale. From the instructors' standpoint, as the Guardian's sources make clear, the same system reads as surveillance.

The tension between those two framings — quality control versus monitoring — is the substantive question the story raises for the wider AI-at-work debate. Companies adopting AI management tools argue they standardize output and catch errors. Workers on the receiving end describe stress, self-censorship, and the sense that professional judgment has been replaced by compliance with whatever the model rewards.

Why the stakes extend beyond Multiverse

The Multiverse report lands amid a broader push by UK employers to deploy AI across workforce management — hiring screens, productivity scoring, communication analysis. Teachers and other highly skilled professionals have generally assumed automated monitoring applied mostly to lower-wage, high-throughput roles. The Guardian's findings complicate that assumption: at a £1.6bn firm staffed by skilled instructors, the AI is scoring the delivery of AI skills training itself.

There is a specific irony in the details worth stating plainly: the people being rated by AI are, in many cases, teaching workers how to use AI. The students in these apprenticeship programs are learning the same class of technology that is now generating scores about their instructors.

For policymakers, the case adds a data point to ongoing debates in the UK and EU over regulating AI in the workplace, where employee-monitoring applications have drawn particular scrutiny. For the training sector, it raises a practical question about retention: teaching staff who describe their working conditions in terms of "horrendous stress" and "remorseless" monitoring are staff who may leave, in a market where qualified technical instructors are already scarce.

What the sources actually said

The Guardian's reporting rests on instructors at the company who spoke about conditions internally and to the paper. The key claims, all attributed to those staff, are:

• The AI analyzes transcripts of online classroom sessions and scores teaching performance.
• The system is programmed to alert human managers when instructors are suspected of "doing things wrong."
• Staff report "horrendous stress" and a feeling of being constantly watched.
• Teachers characterized the system as "remorseless" and "unnerving."

The Guardian's report is an exclusive based on whistleblowing from within the company. The full findings, including the company's response to the allegations, appear in the Guardian's continued coverage of the story.

The so-what

The question the Multiverse case puts to every employer deploying AI performance tools is the one the teachers themselves have answered by speaking out: whether a monitoring system that improves operational consistency is worth the cost of a workforce that describes itself as watched, stressed, and unable to trust its own working conditions. If instructor turnover, teaching quality, or regulatory scrutiny follow, Multiverse's £1.6bn valuation will rest partly on how the company answers that.