A little while ago I was at a construction industry seminar on artificial intelligence, and I did something I don’t always do — I put my hand up and spoke.

My point was simple: I wondered whether, as an industry, we might be moving too fast on AI. The speaker didn’t agree. “No,” he said — if anything, Manchester is moving too slow.

He wasn’t wrong, and I want to say that clearly up front. Greater Manchester should absolutely embrace AI. The firms that use it well — to write quicker, plan better and cut out the busywork — will pull ahead of the ones that don’t. Being left behind is a genuine risk.

But that wasn’t quite what I meant. My concern wasn’t the speed of adoption. It was the speed of adoption without any obvious data security measures or policies behind it. In some organisations, that gap is exactly what turns a helpful productivity tool into a serious data risk.

Fast is fine. Fast with no seatbelt isn’t.

Here’s what I see on the ground. Staff are already using AI — often free, public tools — whether the business has sanctioned it or not. They’re pasting in client details, contracts, drawings, tender documents and spreadsheets. It saves them time, and I understand why they do it. But in a lot of firms there is no AI usage policy, no guidance on what can and can’t go into these tools, and no technical controls to stop sensitive data walking out of the door. That is moving too fast.

The same is true when businesses switch on tools like Microsoft 365 Copilot without preparing first. Copilot will happily surface any file a user can technically access — so years of loose folder permissions suddenly become a live data-exposure problem. The tool didn’t create the risk; the rush to deploy it simply exposed one that was already there.

Even the people building AI are saying “pace it”

What struck me this week is that my slightly awkward seminar comment has just been echoed, far more loudly, at the very top of the industry. As Sky News reported, the head of AI company Anthropic, Dario Amodei, has called on the leading “frontier” AI firms to slow the pace at which they push capabilities forward and to let independent experts verify their safety work. His essay is titled, tellingly, We Must Pace the Frontier.

He was quick to say he isn’t calling to halt progress — only to make sure enough time goes into safeguarding these systems. And, remarkably for such a competitive industry, his rivals agreed: Elon Musk replied simply “Dario is right,” and OpenAI’s Sam Altman said the frontier needs pacing too.

If the people building AI are urging caution and guardrails, it seems only sensible that the rest of us — the businesses using it — put our own guardrails in first. Not to slow down, but so we can move fast with confidence.

What “doing it properly” actually looks like

For most Manchester firms this isn’t about science-fiction superintelligence. It’s about a handful of practical, unglamorous basics:

  • A clear AI usage policy — simple rules on what staff can and can’t put into AI tools, written for people, not lawyers.
  • Data loss prevention — controls that stop confidential information being pasted or uploaded into unsanctioned tools.
  • Copilot done right — tightening permissions and applying sensitivity labels before you switch it on, so it only ever surfaces what each person should see.
  • Staff training — because your people are your strongest line of defence once they understand the risk.
  • The fundamentals — multi-factor authentication, monitoring and Cyber Essentials underneath it all.

Get those in place and you can say yes to AI with confidence, rather than crossing your fingers.

So — too fast, or too slow?

Honestly, both. The speaker was right that Manchester shouldn’t drag its feet, and I’d never argue against businesses adopting AI — helping them do exactly that is part of our job. But speed and safety aren’t opposites. The organisations that win with AI won’t be the ones that adopted it fastest; they’ll be the ones that adopted it fastest with the right guardrails in place. That was the point I was fumbling towards at that seminar — and it’s the bit that keeps your data safe.

If you’re rolling out AI and you’re not sure your policies and protections are keeping up, that’s exactly what we help with. Take a look at our approach to AI security, data security and practical AI adoption — or just get in touch for a straight conversation.

Source: Sky News, “AI giants pledge to act over fears tech is developing too fast” (September 2026), reporting on Anthropic CEO Dario Amodei’s essay “We Must Pace the Frontier” and the responses from Elon Musk and Sam Altman. This article reflects public reporting at the time of writing and is offered as general guidance.