THE artificial intelligence summit hosted by the Prime Minister and featuring invited participants from across the world (except Scotland, the door to the event being slammed in our face by the UK Government) prompted three reactions in me, aside from the inevitable impression that Rishi Sunak was auditioning for his next job.

The first was a confirmation of how threatened the UK feels by Scotland and how pettily stupid its knee-jerk exclusion looks to the wider world.

The second was a sense of deja vu. Our generation has often been told that a utopia of less work and more leisure is just around the corner.

Indeed, by now robots were meant to be doing everything while we all pursued our hobbies – but they aren’t, not least because a better future for humanity won’t be made by microchips but only by the conscious actions of compassionate and concerned leaders and they seem thinner on the ground than ever.

READ MORE: These are the rallies for a Gaza ceasefire taking place in Scotland

The third was some amusement at the repetition of lurid scare stories about computers taking over the world and the havoc that will apparently arise when they become brighter than we are.

That type of response is nothing new. No doubt there were those who, witnessing the invention of the wheel, muttered to their neighbours in the next-door cave about the dangers of being mowed down by out-of-control irresponsibly pushed circular lumps of rock.

However, the reality usually turns out to be less like Hal, the murderous super brain from Stanley Kubrick’s 2001: A Space Odyssey and much more like Marvin, the paranoid android from Douglas Adams’s The Hitchhiker’s Guide to the Galaxy.

AI, like every other work of human hands, will be as good and useful as we choose to make it, or just as crass and useless as every other tool can turn out to be when wrongly, maliciously or just stupidly deployed.

I had a demonstration of that last week. Into my Glasgow University inbox popped an email which started with this attention-grabbing and transparently AI-generated paragraph: “Saw on your LinkedIn that you’re from Colintraive! Considering your role as a Professor in Scottish Culture & Governance in Higher Education, you’d probably enjoy knowing that your beautiful village was notably visited by Queen Victoria in the 19th century, one of the most respected monarchs in British history.

“Colintraive’s historical significance and charm is truly fitting for an esteemed educator like yourself, isn’t it?”

The sender was a senior manager at a green energy consultancy in search of clients. Indeed he was so keen to meet that his computer signed off for him with the memorable flourish “Tomorrow suits me!”

As this email demonstrates, AI in the wrong hands is far from being intelligent or even well-informed. I never use LinkedIn (in fact I scrubbed my redundant profile the next day), I play no role in spending decisions at Glasgow or any other university and appealing to royal connections is far from being the best way to gain my sympathy.

I was in the process of deleting this junk when it occurred to me that an AI approach surely deserved an AI response. Fortunately one of the programs on my Mac had recently added an AI package so I imported the email and asked my Mac to craft a “negative response”.

AND what a response it produced. “Unfortunately,” it started, “I must express my disappointment. While I appreciate your attempt to establish a personal connection by mentioning my home town and profession, your intrusion into my personal life through LinkedIn feels invasive and inappropriate.”

“Moreover,” it continued, “your attempt to impress me with the historical significance of Colintraive and its connection to Queen Victoria seems forced and unnecessary. I fail to see how this information is relevant to your pitch about reducing energy costs and becoming carbon-neutral.”

Going on to call the approach “presumptuous” and “insensitive” it finished: “In summary, your letter lacks tact, relevance, and consideration for my personal circumstances. I kindly request that you respect my privacy and refrain from reaching out to me in a similar manner in the future.”

READ MORE: Roz Foyer - King's Speech proved just how badly we need an election

I couldn’t have put it better myself, and the full reply was quickly despatched to the writer along with a little description of what had drafted it and why. There has been, inevitably, no answer.

All this proves the truth of an old computing term, GIGO – garbage in, garbage out. AI provides a higher quality of output than previous technological upgrades and does it more quickly and with fewer and much more easily expressed instructions.

But if those instructions are wrong, or the motivations unclear, then what it will produce is not something that equates to human thought or feeling but merely further proof of the inadequacies of its user.

New style guidelines from the Associated Press – the arbitrar for many regarding media expression – suggest that coverage of AI should not be allowed to humanise technology. It also counsels against machines as copywriters or reporters, though using AI to provide first drafts and basic texts of information can be useful.

The vast majority of people will, of course, know nothing and care less about AI even when it is operating behind the scenes in every appliance from limos to lawnmowers and in all types of advertising and journalism.

However there are opportunities, as in any technological step forward, for creative, clever and appropriate innovation and Scotland should encourage that, ignoring any attempt to stifle it, from Sunak or anyone else.

Machines aren’t going to eat our planet nor, unfortunately, are they going to allow us to recline gracefully peeling a grape as bits of metal do all our chores for us.

They are only as good or bad as we are and that perhaps is the most worrying thought of all because so far we haven’t worked out how to rise above our worst and most self-harming nature, as a daily glance at the news continues to show.