Scotland needs to embrace artificial intelligence and trust its own planning system if the technology’s potential is to be realised, innovation minister Alyn Smith has said.
The Scottish government, citing independent research, says AI could generate an additional £23bn in annual GDP for Scotland by 2035 with the right investment and leadership. The figure arrives in the middle of an argument that is not really about AI at all, but about where the buildings that run it are allowed to go.
The planning fight
In September the Scottish Greens led a debate calling for a pause on planning applications for large-scale AI data centres until national guidance was published. The government rejected that formally, while saying consent for such sites was “highly unlikely” before guidance is updated by the end of this year.
Ministers acknowledge “legitimate concerns” about large data centres and say oversight of the planning process is being increased. People living near proposed sites have told BBC Scotland they worry about how the developments will look, the noise, and the amount of land required.
Experts have warned MSPs that the centres could push up electricity prices for consumers and make climate targets harder to meet. The Scottish Greens argue that unregulated expansion would “enrich giant AI companies and their super-rich owners”.
What the minister actually said
Smith was blunt about the tone of the debate. “Some of the discourse we have seen on this is wild,” he told Radio Scotland Breakfast. “You’d think there wasn’t a planning system at all. Comparing a really bad data centre in the middle of the Utah desert doesn’t track through to what we’re going to have in Scotland.”
He reached for a comparison to make the scale argument: the IT industry in Scotland currently employs 87,000 people, against the 21,000 employed by the mining industry in 1979. “Lets just get a grip on this,” he said. “The people of Scotland are serious about technological developments.”
On the centres themselves he used a Scottish metaphor, calling data centres “the whisky bonds of the AI sector” — warehouses holding something valuable while it matures. “If you’re using a smartphone you ask where your data is being kept. Is it somewhere reliable? Is it somewhere trustworthy?”
The admission in the middle
The most useful sentence Smith offered was a concession: he admitted data centres do not create a great many jobs.
That matters, because jobs are the argument usually made for them locally. If the case is not employment, it has to be something else — tax base, grid investment, sovereignty over where national data physically sits. Smith leaned on the last of those, and on benefits to businesses that use AI rather than host it.
He also said the planning system would decide case by case: “Not all of these applications are going to go forward because not all of them are suitable or have sufficient community engagement.”
Skills, jobs and guard rails
Smith was due at a meeting in Edinburgh to hear from businesses about what is helping and hindering AI adoption, with the role of Scotland’s universities and colleges in supplying AI and data skills also on the agenda.
He pointed to a new Workplace Jobs Observatory, led by the STUC, which could examine in real time how AI is affecting work. And he set out where he wants regulation to land: “I want to see ethical guard rails put on it by legislation and want to align with the EU norms.”
Why the number deserves scepticism
£23bn a year by 2035 is a projection of what a technology might add if adoption, investment and leadership all go well. It is not a forecast and it is not a commitment, and figures of this shape have a habit of being quoted later as though they were either.
The more testable claims are the ones with dates attached: guidance by the end of this year, and consents unlikely before it exists.
