In recent years, CEO James Cobb started to worry that artificial intelligence (AI) was being used by employees everywhere in his company. From Slack messages to emails, documents, internal updates – all of it. And, as a result, he worried that he was no longer spending his work days talking to other humans. “I don’t want to end up with my AI agent talking to someone else’s AI agent. If I want an update on what’s happening, I want to hear from you. I don’t want to hear AI,” says Guildford-based Cobb, who’s the CEO and founder of Crowd Connected, a location technology company. “I think the assumption [by users] is that AI is going to write it better than you could, and it doesn’t. AI is bizarrely bad at English. It actually can obscure the underlying message, or even distort it and get it wrong,” he explains. Shorts So in February, he introduced a company-wide policy – no AI on internal communications so his team would use their “brains and voices”. More than six months later, he admits to perhaps being too generous in permitting people to still use AI to structure points and check their final message because the change hasn’t worked: AI is still noticeable in the emails and documents he’s receiving. “If you allow AI to write something and then think you’re going to use that as a starting point and human‑edit it, it never ends up that way, because AI’s fingerprints are always there,” says Cobb. “Maybe we have to say, ‘This isn’t working. We have to stop using it now’.” The “Botshitting” backlash Cobb is right that AI seems to have infiltrated all areas of work. 90 per cent of UK digital workers use AI at work, according to the 2026 Work AI Index by Glean – a platform for using AI at work. Only 18 per cent of these say it’s significantly improved their organisation’s performance, largely thanks to the two new trends of “botsitting” and “Botshitting”. “Botsitting” is the time workers are increasingly spending making AI generated content usable – UK workers spend about six hours a week on this, according to Glean. “Botshitting”, meanwhile, is when workers distribute AI generated work without checking they actually understand it themselves – the next evolution of AI slop. Worryingly, 70 per cent of UK AI users in Glean’s survey admit to doing this. The stats justify Cobb’s desire to distance his business from AI, despite the prevailing business message that those who don’t embrace AI will be left behind. Cobb and other business leaders are now questioning its usage and its value, with 51 per cent of business leaders expressing concern about employees using unapproved AI tools, according to TrustedTech. And the worst culprits for AI overuse, says Cobb, are his Gen Z employees – who make up half of his 10-strong team. While he believes this stems partially from being “AI natives” – and perhaps overconfident in AI’s abilities – he senses there’s something deeper at play. “They’re not confident that they’re talking in the right way. They think there’s some magic business way to talk, and that they need to write updates in the right way – that AI is going to give them the gloss to make them look clever,” says Cobb. ‘The expectation should be simple: if your name is on the work, you must understand it, check it and take responsibility for it,’ says AI researcher Kamila Miller Kamila Miller, an applied AI researcher and lecturer at Henley Business School, says that far from making them look clever, the reason Gen Z rely on AI is because bosses are encouraging its usage. Miller cites data from the Pulse of Work 2026 survey that said 76 per cent of Gen Z feel pressured to use AI – against 60 per cent of people overall. And 46 per cent acknowledge that relying on AI is making them less intelligent. “Gen Z are not the problem. They feel the pressure most, have the least experience to fall back on and are the first to notice their own skills slipping,” says Miller. “Blaming them for responding to incentives their employers created is not a policy. Slop is not an accident. It is what you get when pressure to produce meets a tool that makes producing effortless.” Over at Hype Presentations – which is introducing a “No AI Day” this month – it’s the opposite problem; older team members are the most guilty of using AI to draft their emails and documents, admits Manchester-based Chris Tomlin, the company’s co-founder. “The elder team members – myself included – were looking at AI whereas the younger team weren’t sold from the beginning.” Now Tomlin and his leadership team have become concerned that AI is overriding people’s ability to brainstorm, problem solve and even write. He’s spotted obvious AI “tells” in emails from team members – certain phrases, em dashes or sentences written in sets of three – and doesn’t think it’s fair for clients to receive emails like that. “If I can see it, other people can see it. If you cared enough, you’d write the email yourself,” says Tomlin. He is introducing a “No AI Day” for his workers to serve as a reality check for just how often everyone in the office is turning to AI – and rethink whether they really need to. “It can be quite easy to say as a default, ‘I’ll pop that into Claude’, or ‘I’ll pop that into Copilot’, so it will be interesting to see how many times we attempt to do that in a day,” says Tomlin. He wants to use the day as a barometer to show people what they can achieve without relying on AI. “It’s a bit of a pressure test really – we don’t want to become over-reliant on AI,” he adds. While Miller agrees that AI slop and “Botshitting” is unhelpful and unproductive, and that initiatives like “No AI Day” could be useful, she thinks an AI ban isn’t a convincing policy. “A ban will mostly produce people quietly using AI on their phones… the expectation should be simple: if your name is on the work, you must understand it, check it and take responsibility for it. If you cannot explain it, you did not do it,” she says. For Cobb, he says until models can write English well, it is time for his employees to pause using it all together.
I banned my employees from using AI – they think it makes them look clever
Full Article
Original Source
Read the full article at Inews →KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.