When does AI stop being AI? When it starts being useful.
By Peter Springett · 29 September 2026
I speak about AI to lots of people. At work, with friends, just chatting with people in shops or at the run club. There’s a lot of anxiety. Some of them are afraid that AI will kill us all, are worried about the environment, others that they or a family member will lose their job.
The other thing they all have in common? They’re using AI at home, not just at work. The golfer who uses ChatGPT to analyse their swing, the novelist using it to manage savings, or a family getting availability updates for the ideal skiing holiday, tailored to age, ability and budget.
I mention this because most articles that I read these days make a binary distinction between the AI-pilled and AI refuseniks. Depending on the author, we are on the brink of an era of superabundance, or we are on the brink of the apocalypse; data centres will lead to water shortages, or AI will build us a new energy grid; and so on.
There’s little or nothing for the AI-normies. People who are anxious about the future but in the next breath using LLMs to get on top of their pension, their handicap and find suitable nursery slopes for the kids.
Meta makes a mark with Muse
That's why I think Meta is on to something with Muse, its new AI assistant. Rebranding AI at home by making it fun as well as useful is a smart move. It’s the start of a process where we might even stop saying AI and call it Muse or Siri instead.
Contrast this with the workplace, where AI is still AI and where imagination takes second place to the generic. Deloitte UK recently found the top workplace uses are information search, email drafting and summarising documents. Dull work made worse by limited training and unclear leadership.
One area where it is making an impact is in developer teams. AI is extremely good at coding and debugging. But this applies to just 8% of workers says the ECB. Most of us are still just using AI to generate our own emails and condense those of our colleagues. We’re digging holes and filling them in at the same time. AI is both slog and slop.
Not surprising given the lack of direction. According to a recent survey from Culture Amp, 85% of employees are told to use AI. Yet 42% say leaders have never explained how it can support company goals. If you can’t say what AI is for, it doesn’t matter how good the platform is or the size of your investment. People won’t get on board.
AI: Home for the holidays
I think that seasonal events are driving popularity as well. When people get together to show off new devices and new apps adoption kicks in. The summer holidays ended a month ago and I’ve definitely seen an uptick in AI use since then. Thanksgiving is in two months’ time. That’s plenty of space for Muse to start gaining traction in the US and for people to have something to talk about around the turkey.
Same for the UK at Christmas. When people compare new smartphones, it won't be the cameras. They'll be demonstrating Muse, Siri and the voice editions of ChatGPT and Claude: Moving home, getting a flight refund, fixing a golf swing. The more confident will be bringing along their own apps to play favourite games or set a kitchen schedule for vegetables, gravy and sausages.
From OpenClaw to Muse in nine months
Take a step back and it’s been an incredible year. In January 2026, OpenClaw arrived with a mixture of excitement and terror. As long as you had a day to set it up, it could triage and answer your emails. And sometimes delete them without you asking. Then Cowork and ChatGPT work proved that agents could run safely and plug into a multitude of apps.
Now we have Muse and a revamped Siri surging in popularity because at home nobody needs a training programme or gets told what model to use. They see a problem and fix it in less than ten minutes. That's the lesson for the workplace. Cowork, Copilot and whatever Gemini brings next won't break the email treadmill on their own. People adopt AI when someone shows them what it's for. The kitchen table has worked that out. The boardroom hasn't. Yet.
Sources:
Deloitte UK: GenAI Workforce Survey, published 16 September 2026
ECB: AI adoption and the productivity promise: what workers report, published 26 August 2026
Culture Amp: AI at Work Benchmark AI at Work Benchmark, published 17 September 2026