From the arguments I’ve witnessed, it seems on one side there are people who treat AI tools such as OpenAI’s ChatGPT and Anthropic’s Claude like a helpful colleague, and on the other, those who view it as the professional equivalent of sneaking a smartphone into a school exam.
It doesn’t help that our workplaces aren’t entirely clear on where they stand either. In the UK, only 54 per cent of firms actively use AI, according to the British Chambers of Commerce. Yet last month, Deloitte surveyed 25,000 UK adults and found that Britons spend £1bn a year of their own money on AI tools for work, and a third of people are using it without their employer’s permission.
The AI divide is having effects on office life when it comes to communication, management and job security. One former colleague now in the publishing industry tells me her pet peeve is when people send messages written by AI – which she described as “disrespectful” because it indicates they don’t deign her worthy of real, human interaction.
Meanwhile, at the finance company where my friend works, the software engineers almost entirely depend on AI to communicate with one another – via their instant messaging platform, they can ask it to draft a message based on notes, or hand it a task such as summarising what has been said.
It’s the new tension over work communication styles – do you value personal attention more than efficiency? Does a message feel less genuine because an AI helped write it? Some people would rather spend a few minutes crafting a considerate message, for fear of coming across rude. Others are happy to risk being thought of as brusque because of time constraints. (All that said, there may be a middle ground – where, say, you prompt AI to craft a message, and then refine it yourself.)
The divide can go so far as to cause rebellion, especially when people have strongly held principles, caused by bigger, existential fears around livelihood and identity. At a tutoring company an acquaintance works at, one manager has raised objections to being encouraged to use AI, and turned the features off on her laptop, arguing that she’s training an artificial intelligence to do her job. Her boss and company haven’t done anything about it yet, but the issue isn’t going away and her anxieties aren’t unfounded. UK companies surveyed by Morgan Stanley earlier this year reported an eight per cent per cent net reduction in jobs attributable to AI over the previous 12 months – the highest rate among major global economies.
Arguably, however, the biggest divide is over the ability to delegate and let go. I have heard colleagues and friends say that they don’t use AI, since they tried asking it to do an aspect of their job and it went terribly. Another will shrug and respond that this is to be expected: the AI needs feedback to learn, and time to become adept at the task.
It reminds me of a former boss who was always overloaded, yet wouldn’t spare the time up front to teach the team to take things off their plate. We all know how hard it can be to delegate at first, for fear of the task not being done exactly as we’d like. But eventually we relinquish control because the reduction in quality may be worth the time it gives us back to do the things we’re best at. Like AI, humans make mistakes too. We seem willing to forgive errors in a colleague, while treating the same slip-up by the AI as proof it cannot be trusted.
Indeed, I’m not immune to anti-AI instincts myself. When people I knew first started talking about paying to subscribe to Chat GPT a year ago, I was dismissive and judgemental. I cited all kinds of concerns around overreliance and data privacy, but really it was about trying to maintain perfectionism. Now I pay £20 a month for a Chat GPT Plus subscription, and use it to identify how to fix a tap or light that’s stopped working, troubleshoot my tennis technique, and sync it with my Google calendar so everything is in my diary.
AI is already becoming part of working life, whether we welcome it or not. Learning how to use it well could become as fundamental to our careers as learning to use the internet. But the skill may be in deciding which tasks to hand over, and knowing how to check the results. Using AI to summarise a long document, organise notes or cut down on admin can free us up for work that needs our judgement. Relying on it to design a bridge or preside over executive decisions might be less sensible.
But those who refuse to engage with AI at all risk being left behind, as colleagues learn what these tools can do. We can opt out of using it, but in all likelihood, we can’t opt out of a workplace being changed by it. You don’t have to love AI. But refusing to learn how to use it could soon cost you.