A government minister has confirmed that Scotland’s AI plan is being ‘refreshed’ during a parliamentary session focused on regulation for the technology.
Richard Lochhead, business minister, committed to taking ‘on board’ considerations about the risks that surround the technology during portfolio questions at Holyrood yesterday.
Mr Lochhead, who has previously said Scotland is ‘open for business‘ on AI, was answering a question from Greens’ MSP Patrick Harvie about there being great benefits from its implementation, but also potentially ‘catastrophic risks’.
He called therefore for a regulatory approach that seeks to balance those risks and benefits in a Scottish context, when legislatively the technology is reserved to Westminster, and there is also divergence from the EU.
He asked specifically, though, whether a risk assessment had been carried out by the Scottish Government, when it will publish any risks that have been identified, how they might be managed and what the implications are for regulation.
Mr Lochhead said: “Patrick Harvie raises a number of important points. First and foremost is the point about the balance between addressing the risks of AI and capturing its massive benefits at the same time. That should be reflected in regulation.
“As he says, it is also important that we guard against divergence from EU regulation, given that many of the companies in this country export to the EU. We have a UK Labour Government that is making targeted interventions to address the AI risks, whereas the EU has taken a much more comprehensive approach to the issue. I assure Patrick Harvie that we in Scotland are refreshing our AI plan and we will take his valid points on board.”
Scotland published its own National AI Strategy in March 2021, based on a commitment to use the technology on it being ‘trustworthy, ethical and inclusive’. However the strategy focused more on the rights-based approach to the technology than specific risks. It did however mention the risk of bias in training algorithms, rather than broader societal risks that have been raised since the advent of generative AI, which went mainstream with the launch of ChatGPT in 2023.
Mr Harvie added that the Scottish Government’s AI strategy has ‘little to say about risk, but it endorses one of the Organisation for Economic Co-operation and Development’s principles, which is that potential risks should be continually assessed and managed.’
In a recent question to the Scottish Government about whether its staff would be able to use the new generative AI platform DeepSeek, developed in China, Futurescot was told that the Scottish Government’s AI policy adheres to the framework published by the UK Government which sets out 10 principles to guide the safe, responsible and effective use of generative AI in government organisations.
During the session at Holyrood, Liz Smith of the Scottish Conservatives, also asked about the implications for AI in misinformation, and the technology’s use to potentially undermine elections.
With an election at Holyrood itself looming next year, Mr Lochhead said he has raised the matter ‘time and again’ with the UK Government.
He said: “One of the issues of concern to the Scottish Government is that, although the UK has tended to talk about the long-term frontier threats of AI, that has come at the expense of talking about and addressing the short-term risks of misinformation, deepfake images, or other more immediate threats that the public and business community, which want to use AI in a way that can be trusted, are very concerned about. That is a point that we are conveying to the UK Government.”
However, Mr Lochhead’s call for a ‘four nations summit’ in 2023 – involving the UK Governnent and the devolved administrations – went unheeded.
A Scottish Government spokesperson said: “We announced in June 2023 that the Scottish AI Alliance would be commissioned to lead an independent review of Scotland’s AI Strategy, making sure it evolves and keeps up with the accelerating pace of change in AI. The Alliance carried out that work across the second half of 2023 and 2024, and we are now at the stage of considering that review with a view publishing a refreshed strategy later this year.”