What happens when a director from Sony Music sings Bob Dylan, a professional composer challenges the use of AI, and a room full of people starts questioning what creativity actually means?
At Digital Somerset’s DS22: CTRL + CREATE, we explored AI-generated music, copyright, human creativity and the importance of making mistakes. Along the way, we discussed cooking, the shark from Jaws, and why we might all need a gym for our brains.
An evening of different opinions, excellent audience questions and plenty of laughs left us with much more to think about than we’d expected.
AI can produce a 300-page legal argument in minutes, complete with convincing language and references to cases that might not even exist. So what happens when people start trusting the appearance of expertise without understanding the work behind it?
Drawing on recent cases in British courts, new guidance for solicitors, and an analogy from Stephens Scown’s Thomas Chartres-Moore, I explore how AI is changing the relationship between clients and legal professionals, why helping people ask better questions is a positive step, and why human judgement and responsibility matter more than ever.
Because just because AI makes something look and sound right doesn’t mean it actually is.
Last night was another Digital Hub, and I came away from this one thinking we’d managed to hit quite an interesting point in the AI conversation.
Every couple of months I stand up there and try to give some kind of update on what has happened since the last event, and that job is getting increasingly ridiculous. I normally start building the slides a week or so beforehand, trying to work out which developments actually matter and which ones are just noise, and this time the whole thing had pretty much changed again by the time I stood up to deliver it.
This is the first in a series of pieces exploring our relationship with technology: what it gives us, what it costs us, what we choose to hand over, and what we consciously choose to keep for ourselves. The wider thread is something I’ve been calling Do More Human Work, although the thinking reaches well beyond work itself.
What technology gives us, what it takes from us, and whether we’re still conscious enough to choose the trade-off.
The people building the world’s most powerful AI systems are starting to ask whether we are moving too quickly. I think we should listen. I also think we need to be very careful about which AI we’re talking about.
This morning, just after 7.00, I found myself on BBC Radio Somerset with Charlie Taylor being asked a fairly cheerful breakfast-time question: how concerned should we be that AI could kill us all?
Smart glasses have become a lightning rod for arguments about privacy, surveillance, accessibility and trust. In this piece, I look past the headlines to what Meta’s glasses actually do, where the genuine risks sit, how people are already using them for independence and accessibility, and why blanket bans may end up punishing the wrong people. Since first publishing the article on LinkedIn, the debate has widened further through venue restrictions, media coverage and strong public reactions, but the same question keeps coming back: who do we trust, the technology, the companies behind it, or the people wearing it?
Creatives can really embrace AI as a tool for pushing boundaries and enhancing their unique ideas rather than how others are using it merely as a shortcut.
Scott Quilter reflects on his keynote at Somerset Business Expo 2026, covering AI prompts, agents, skills, co-work, and why human judgement still matters.
The viral critique of a Claude Monet painting mistakenly labeled as AI-generated highlights how preconceived perceptions can overshadow actual quality in art and other professional evaluations.
New AI models, Mythos and Daybreak, enhance cybersecurity by changing the economics of vulnerability research, but their capabilities may not be as groundbreaking as initially claimed.
This week’s AI Roundup comes from a sunny office, a packed calendar, and a first talk as FBCS at Bath Digital Festival. I’ve written up the session, including the live demos, the ROAR framework and the 4 rules we keep behind every serious AI conversation. There’s also a new piece on why buying licences is miles away from having an AI strategy, a closer look at what Gemini, Claude and ChatGPT do with your conversations and feedback, plus the usual run through the week’s AI news: Google I/O, ChatGPT in PowerPoint, Microsoft’s multi-model Copilot work, AI security agents, provenance standards, and the quiet end of Ask.com. Add in a Monet mistaken for AI, a practical handoff prompt for long chats, the return of Prompt Fiction, and a few diary updates, and it has been one of those weeks where the industry feels loud, messy, fast, and genuinely useful if you know where to look.
ChatGPT’s unexpected mentions of goblins and other creatures stemmed from an incentive misalignment during its training, leading to a widespread issue that OpenAI needed to address.
The AI Roundup welcomes new subscribers and highlights the growing demand for practical AI and automation training, while also addressing recent developments and challenges in the AI landscape.
Details from a recent supported workshop utilising a custom AI bot helped diverse participants clarify their business problems and create actionable plans by focusing on structured problem-solving, and how this looks like it could be the future for education.
I wrote a version of this piece in March and never published it. Why? The world was moving fast enough that anything I filed would have aged badly within a fortnight. That instinct turned out to be right. Because what’s happened since then is one of the stranger business stories of the year, and you’d miss the full picture reading the headlines as they arrive.
So here’s the whole arc, in one place.
Mel Robbins’ Copilot backlash is about more than one bad prompt. It exposes a bigger problem with AI advice, data safety, and the lazy habit of treating women as if they are somehow “behind” on a technology that has already become mainstream.