Computer says d'oh: Hours at work are spent fixing AI mistakes

Published: 18:53 EDT, 4 November 2026 | Updated: 20:35 EDT, 4 November 2026 The growing use of AI in UK workplaces is creating extra tasks for employees.On average, workers spend nearly seven hours a week fixing AI–generated content, including removing unnecessary waffle and correcting incorrect facts and figures. Almost half have had to correct or redo inaccurate work produced by colleagues using AI.A study of 2,000 UK workers, commissioned by Arden University, found that 64 per cent have used AI at work in the past three months, mainly for research, writing emails and generating ideas.The study also found that 19 per cent have had to completely redo a colleague's AI–generated work and that 53 per cent do not tell management when they use AI.Some 39 per cent are unsure which AI tools they are allowed to use and only 13 per cent feel confident identifying fake sources or references.Meanwhile, 84 per cent of AI users say it saves them time overall, despite the need for corrections. Younger employees are more likely to use AI, with Gen Z (73 per cent) and Millennials (72 per cent) ahead of Gen X (56 per cent). A study found that almost half of UK workers are spending nearly seven hours a week correcting or re doing AI generated workArden University says better AI training is needed so employees can give clear instructions, check accuracy and use the technology responsibly. Vitaly Klopot, chief innovation and commercial officer at Arden University, said: 'Getting an answer from a chatbot is only the starting point. AI literacy means choosing suitable tasks, giving clear instructions, checking answers and understanding what information you can safely share.'The opportunity goes beyond producing more emails or documents. It's about solving problems, improving work and building reliable processes for everyday tasks.'A quick draft may save one person time, but if a colleague spends hours correcting it, employers need to consider the effect on the whole team.'Practical education can help workers move beyond trial and error. Most users already report saving time – the next step is helping more people get useful results confidently and responsibly.'