OpenAI, the US-based artificial intelligence (AI) company, recently unveiled GPT-5. Described as ‘PhD-level’ in coding and writing, ChatGPT features are set to roll out across Apple devices via Apple Intelligence later this year. To be clear, ChatGPT is only one aspect of AI – not the discipline itself – and one of many platforms, albeit the most familiar today.
With a top five spot on the Microsoft Store, top ten on Google Play for Android, and reigning supreme at number one on Apple’s App Store, it comes as no surprise that the company is on track to reach around 700 million weekly users. Taken together, it would be fair to say that ChatGPT certainly isn’t going anywhere, it’s just getting started.
‘Digital Lead’ may sit proudly in my job title and imply I’m a tech pro, but the truth is that when it comes to coding, data tools and engineering, or AI, I’m in over my head. To put that into context, here are genuine questions I’ve thrown at ChatGPT:
- ‘If I climb a Munro, will there be midges?’
- ‘What is the highest-calorie meal to ever be created?’
- ‘Create a “wanted” image but it’s a picture of a black North Face slider’ (Sadly, still lost).
Evidently, AI hasn’t been my go-to for personal tasks, yet as the topic dominates headlines and debates worldwide, its capabilities are becoming impossible to ignore.
Behind those humiliating questions lies a steep learning curve. I now find myself knee-deep in ChatGPT tutorials to get up to speed. Critics will say that AI makes us lazy, harms the environment or threatens humanity. For me, the real revelation has been uncovering the extent of AI’s capabilities – but only when guided by human judgement.
The best advice I’ve had so far comes from the computer science rule: ‘GIGO’, meaning ‘garbage in, garbage out’. Much like the human body needs good nutrition to stay fit and healthy, AI systems require high-quality data and clear humaninstruction – feed either of them junk and you’ll get junk back. (Or think of the song “You Get What You Give” by The New Radicals. Same principle.)
Human inputs, or prompts, are what allow us to gather information from AI systems like ChatGPT in the first place. IBM describes this process, known as prompt engineering, as: ‘the process of writing, refining and optimising inputs to encourage generative AI systems to create specific, high-quality outputs.’
Prompt engineering is a skill, one that takes patience and a lot of trial and error to develop. I like to think of a prompt as a mini project: you set its role, outline the task, share the context and specify the output format, then ask it to self-critique and improve using iteration methods (tweak, test, repeat). The process is iterative. You find yourself working alongside AI to get the answer you need: refining prompts, introducing constraints, providing extra context, altering the framework and more.
Now you might be thinking, have I used ChatGPT to write this blog? The short answer is yes, but its role and contribution are more nuanced than it seems. I was born in 2001; I’ve grown up with Google at my fingertips – I’ve leaned on it for everything from thesaurus entries and grammar checks to paragraph prompts and quick facts. As one of the first generations raised on smartphones, instant access to online tools has become second nature. The only difference now is that ChatGPT brings all those tools together.
A quick parallel: handheld calculators were once branded ‘cheating’, yet they are now part of everyday life, taught in the curriculum, used at work and at home, and on every smartphone. Whether an exam allows them or not, students must still show their working. Technology moves fast and AI is quickly becoming an everyday tool, so the same logic should apply. It is more powerful, but with clear rules and transparency it can support rather than replace understanding and judgement.
To reiterate, ChatGPT did not write this blog for me, but I have used it to rearrange sentences, check for repetition, critique my tone and suggest alternatives. Essentially, I’m working with ChatGPT to draft, critique and revise my own writing.
No matter where you stand on the spectrum of AI scepticism, there’s no denying AI will be transformational for society. From how we work, learn, interact, travel and access healthcare and much more; the potential for AI feels almost limitless.
From a small think tank’s perspective, it’s about deepening understanding, testing ideas and sharing results transparently. Could feeding it different policy scenarios generate viable solutions? Can it analyse and visualise relevant data in new, insightful ways? What blind spots or alternative perspectives are we missing?
Not forgetting the real-world impacts: as AI adoption grows, so do its energy needs. Every question to ChatGPT runs on servers in large data centres that require significant electricity and cooling – some analyses compare a short chat to using the equivalent of a small bottle of water. There are huge opportunities for Scotland to grow AI and data centres, but that will demand substantially more clean, reliable electricity. The more we use, and rely on, AI platforms like ChatGPT – the more we need an open debate about how to power them.
Ultimately, it’s not about quick fixes or a single plan; it’s about recognising AI is here to stay and opening the conversation about what it means for our work, Scotland’s economy, and the wider policy debate.
Morgan Higgins is Enlighten’s Digital Engagement Lead
1 comment
Steve Bastiman
Scotland’s ambitions in AI and data centre development have reached a critical juncture, but it remains uncertain whether efforts to address this are too little, too late. The country has significantly lagged behind in AI technologies, from wafer fabs to advanced prompt structures, raising concerns about its competitiveness in the field.
For over four decades, AI has been a crucial concern in the business world. However, the current government’s focus on independence appears to have overshadowed initiatives to grasp the trajectory of AI development. As a result, Scotland has missed the opportunity to develop a deep understanding of AI’s direction and technical intricacies, hindering its capacity to keep pace with advancements, particularly in large language models (LLMs), over the past eight years.
While 10 Scottish universities offer AI courses, many of these programs are theory-driven or focus on vague concepts, such as ethical AI. Moreover, few businesses in Scotland are willing or able to pay the premium salaries required to attract top AI talent.
The Scottish Government’s current energy plans and approach have raised doubts about the feasibility of supporting large data centres. The substantial demands of these centres may not be guaranteed, posing a significant challenge to Scotland’s AI and data centre development ambitions.
Key Takeaways:
Scotland’s AI and data centre development ambitions face significant challenges.
The country’s lag in AI technologies, lack of understanding of AI’s direction, and limited AI talent hinder its competitiveness.
The Scottish Government’s energy plans may not support the demands of large data centres.
Recommendations:
Develop a comprehensive strategy to address Scotland’s AI and data centre development challenges.
Invest in AI education and training programs that focus on practical applications.
Encourage businesses to invest in AI talent and infrastructure.
By addressing these challenges and implementing a comprehensive strategy, Scotland may be able to overcome its current limitations, but Government will have to play its’ part IF Scotland is to become a player in AI and data centre provision.