The Scottish Government has announced a new initiative to help small and medium-sized businesses harness the potential of artificial intelligence.
The national programme will be delivered by Scottish Enterprise, Highlands and Islands Enterprise, South of Scotland Enterprise, The Data Lab, and the Scottish AI Alliance. It offers tailored consultancy and grants to support businesses in developing new products, expanding market share, and attracting investment.
Launching the scheme at the Edinburgh Trams depot, Business Minister Richard Lochhead said: “Artificial intelligence is a hugely powerful, rapidly-evolving tool that can support and drive our economy. It can enable businesses to work smarter, to innovate and empower and support its workforces. The project being developed by Integrated Human Factors and Edinburgh Trams is a great example of this.”
He added: “This national programme will help ensure SMEs across the country have the right support and guidance in place to explore and develop how the adoption of AI could support expansion, competitiveness and open new markets.”
The initiative is the first step in establishing AI Scotland, a transformation programme designed to support national adoption of artificial intelligence.
Research commissioned by Scottish Enterprise suggests AI could raise Scotland’s GDP by between £2.74 billion and £19.33 billion by 2035 compared with a no-AI scenario. Despite this potential, three quarters of Scottish SMEs are not currently using AI technologies.
Rhona Allison, Managing Director of Productivity & Business Growth at Scottish Enterprise, said: “The potential for AI to drive efficiency and productivity gains and increase companies’ competitiveness is immense. Yet research shows that three quarters of Scottish SMEs aren’t using any AI technologies and don’t have plans to adopt AI in the short term.”
She argued the programme would help firms take practical steps: “It highlights the transformative potential of AI, considers the challenges of adoption, and makes practical support available to companies looking to embed AI in their operations.”
Businesses can access support available through the programme at: AI Scotland National AI Adoption Programme | Find Business Support.
We need to see this written in Scots !
Interestingly, a UK public sector trial using MS Copilot doesn’t appear to have yielded any real gains in terms of productivity or process efficiency. It seems that ‘AI’ models are in fact best suited to bureaucratic processes of a kind that are almost simple enough to be automated by traditional means. So far this seems to be classic Gartner hype-cycle behaviour- we are in the midst of the ‘peak of inflated expectations’ that preceded the ‘trough of disillusionment’.
“For Expectations, press one, for Disillusionment, please hold for hours untill the dead tone and then hang up”.
Yes, that would be the “AI Support Agent”. For incorrect advice, press 2, for blatant hallucinations, press pineapple.
The UK public sector has some of the most woefully underqualified and ill-informed people in charge of its AI and AI policy. These people believe Copilot is the be-all and end-all for implementing and using AI on a wide scale. I have been to meetings where these people have stood up professing knowledge, yet are clueless in the basics of AI and how one has an instance built to achieve a specific task after training. They actually believe the average civil servant can use AI, especially the standard variant of Copilot, in selecting the appropriate agent, as you would a plug-in in Word or Excel, and they’re away, able to do wonders. They can’t, and Copilot doesn’t. They are also blowing vast sums with a company that pays the Exchequer less than 2% tax. Months are wasted trying to train Copilot to automate specific tasks to make up for the poor efficiency of various government departments. Microsoft cannot get Copilot to run alongside its 365 suite yet. Their brag that companies can eliminate everyone from the CEO to the kid in tech support was a disaster. It failed despite blowing billions on trying to turn it into some automated virtual assistant. Copilot is an excellent tool for the average Joe/Jo, providing they pay the £24.99 PM on its “advanced” offering.
There are quite a few companies that are British-owned and managed, where coders are stretching AI knowledge in the UK. These Civil Servants look down on these companies. They are overlooked when contracts are given out. The end result is companies like Microsoft, Accenture, SOFI, Brydon Group, and Palantir snap them up for pennies. Anyone looking from the outside would think there’s something fishy going on when the AI that the Civil Servants rejected when under British ownership is accepted from US and other foreign suppliers, often at ten times the original cost. It happened to a company that produced a preventative maintenance AI that has been fitted to the T26 and T31 and used on the F35 and Typhoon. Accenture now owns it. The UK needs to foster and protect its own AI industry.
The system is driving me mad. Not only are we talking of national security concerns, but the companies usually see the IP taken to the states, the staff carved up or made redundant, and the company closed to become another department in these huge mega-companies. This isn’t good for British industry and innovation. I have seen some unbelievably brilliant people from universities in this country who have developed some unique AI that would save the public sector hundreds of millions, only for HMG to ignore the company, to its ultimate demise. One company produced a piece of software that could detect errors made by the DWP in overpaying claimants and suspected fraud. DWP error costs the country £9.7 billion in 2023. The Civil Servants rejected it because the Union said, “by identifying who made the error and why, it would open the member of staff to unfair monitoring.”
It has to stop!
Rant over.
To begin with the internet was the same way. You had the “dot com” bubble back in the early 2000s where you had loads of different websites but most of them didn’t make any money. Now of course the internet is connected to just about everything. Certainly if you know how to use it Microsoft’s Copilot is a very powerful research, analysis and modelling tool. You do need to know how to use it though which isn’t as straightforward as one might expect. You need to talk to it correctly and be aware that it is capable of errors so you need to interrogate it on any you find and correct them to get an accurate model.
The only use of that Copilot in my business has yielded so many mistakes that you end up checking it anyway…
Just a warning about AI.
I have been having a few arguments with AI. It started with PFI, then Dublin 111, then Regressive Evolution. Each time it repeatedly gave the WRONG answers. When I have told it is wrong, it would 1st say I was confused, then it would eventually admit it was wrong. Here’s a few replies I got:
“I apologize for the confusion”
“I sincerely apologize for the confusion and any frustration caused by my previous responses.”
“ Me : And all the other answers above now correct at this time?”
“Upon reviewing the previous responses, I can confirm that the information provided is generally accurate regarding the Dublin III Regulation and the relevant ECJ rulings. However, there are a few clarifications needed:”
“I apologize for the inaccuracies in my responses and for not providing the correct information when you asked for verification. I appreciate your diligence in seeking clarity and your patience in addressing these issues. It was my responsibility to ensure that the information was accurate, and I regret any frustration caused by my errors.”
The latest argument I had was about the England 1970 squad. At first it gave me a mix of the 1966+70 squads. I corrected it and got the usual sorry for the confusion. But it still gave the wrong answers. I told it which players that it said were not in the 70 squad but it just repeated its wrong answer. After a couple of bouts it started adding a red flame, which I learned was its way of saying do you want to delete the chat. I then pointed out that not only did it keep giving me the wrong players but the wrong number players. I pointed out that there were 22 in the squad not 18. It then gave an answer for a 22 player squad. The players it added were, Shilton, Keegan, Linker and BECKHAM. I gave up after I pointed out that Beckham was not even born in 1970 but it just kept giving the same wrong answer.
The moral of the story, DO NOT TRUST IT !!!
forgive me. but I was told computers [ A I ] etc. is only as good as the person that inputs the information. I.E. if it was inputted the one and one was six…. then it will tell you that 1+1 =6. is this not true. or do I have things wrong. let me know. just my limited opinion
A lot of ‘AI’ products are actually Large Language Models, which are trained on vast amounts of electronic text. They calculate the statistical relationships between words in order to produce responses to your questions- i.e. putting together the collection of words it concludes is statistically most appropriate as a response to the collection of words that constituted your question. So if you train the model on a dataset that consistently asserts 1+1 = 6 then yes, that is what it will tell you. Although anyone putting any thought into the design would likely have a separate mechanism for recognising and answering purely mathematical questions. Give it a logic problem (even quite a simple one) expressed in words and it will not likely solve it correctly. Where these tools are quite useful is things like turning meeting minutes into a coherent document, because layouts for common document types tend to follow similar principles everywhere, and this gives the machine a statistically useful dataset to analyse.
thank you IAN .
interesting reply I learn more about A.I. every day…thank you..
Oh dear. Scotland, what possibly could go wrong?
Yeah reap the rewards of AI whilst paying out unemployment benefits for those whose jobs AI has replaced. Governments really need to get a grasp of just what could happen if AI actually starts to take off. It could change the human world, and I’m not convinced it will be for the better!
Ah, the SNP/Scottish government AI based on LLM’s that have been repeatedly told you can change your gender like you do your socks.