“I trust my gut feeling more than your AI.” 

This was said to Professor Eyad Elyan back in 2018 when he was hired to use machine learning and artificial intelligence to help solve a company’s problems more efficiently. 

It’s a familiar rhetoric, borne from nascent fears around automation and its implications for the workforce. However, as AI’s use becomes increasingly commonplace, it’s an argument that’s becoming rendered obsolete. 

AI has advanced, and whether we trust it or not, it’s here to stay. 

Profession Eyad Elyan is a Professor at Robort Gordon University, and the lead of the Interactive Machine Vision Research group associated with the university. 

Speaking at ScotSoft 2023, ScotlandIS’s annual conference, he asked an overflowing room ranging from decision-makers to burgeoning tech talent what word came to mind when they heard “artificial intelligence.”

“Robot!” “Speed.” “Terminator!” “Hype.” “Grift.” 

ChatGPT answered “innovation” when Elyan asked it. 

The professor was dissatisfied with each answer. 

“For me, AI is data, it is literally data-driven solutions,” he explained. “And if you want to think about it, you have got to think about data.” 

In fact, whenever Prof Elyan said AI during his presentation at SoftScot, he immediately almost corrected himself to say “data-driven applications.”

With data expanding exponentially, he says, being able to effectively harness this is vital in driving organisations forward, which he says is the main purpose of AI. 

“The metrics are mind blowing,” he said, referring to the recent data boom. 

With exponential data, significant progress at the algorithmic level, and the huge power offered by AI, even home computers have the ability to process, analyse and synthesise different types of data like never before. 

Prof Elyan led the room through the two main forms of AI being used today, and detailed some of the major opportunities – and challenges – the technology offers Scotland. 

Supervised ML – Making Your Data Work for You

Supervised Machine learning is one of the most common types of artificial intelligence used, making up about 95% of the AI landscape today, according to Elyan.

The capabilities and data this technology can harness is completely beyond human ability – “the numbers don’t compare,” claimed Elyan. Ten of thousands for a human, to many millions of AI. 

Supervised ML is highly reliant on quality data, but the capabilities of the type of data it can use has dramatically expanded in recent years. 

Elyan showed a picture of a dog and a child sitting on a porch, with a car driving by in the background. An AI model correctly identified each object in the picture. Just identifying the dog would’ve earned someone a PhD a few years ago, Elyan jokes. 

But the use cases are expanding with the data – complex engineering diagrams and images are simply code and data that AI algorithms can understand, analyse, and make sense of. 

Elyan used this in engineering and monitoring of offshore assets in Aberdeen, reducing workload and providing real-time analytics. But, according to Elyan, the applications are endless, only limited by the amount of data one has. 

He said: “We have lots of data – you don’t have to buy it because your business is generating so much data and you are not using it.

“When we put it to task, it makes things much faster and much better.”

Generative AI – More Questions Than Answers 

Even though it makes up a small fraction of the total AI landscape, generative AI is all the rage. 

Its rise has been prolific, especially since the introduction of ChatGPT.

The large language model, developed by OpenAI, has broken the Turing test, a test developed to see if a machine can exhibit the same degree of intelligence as a human to a believable degree.

“The Turing test stood the test of time for 60 years”, Elyan said. 

“This has been trashed.”

This seminal moment in the technology’s life has created a wealth of opportunities, as well as challenges. 

Challenges and Opportunities

“The great news is that you won’t be exterminated by AI,” Elyan said, however he did leave an ominous admission. 

He added: “There is a price that comes with this, but it’s not going to get rid of us,” he said. 

AI can offer faster, safer and more efficient practices for a number of applications, and it can reduce human bias if the right measures are taken place. 

But there are major challenges that arise from AI’s increased use. 

“Privacy is one of the biggest issues,” Elyan said, as AI data scraping has been a major concern among privacy rights groups. 

Exacerbating human bias is also a worry. 

“AI is simply garbage in, garbage out,” Elyan said. “If you give them biased data, they will give you biased results.” 

With these two issues combined, Elyan says soon, we will miss the simple days of the Cambridge Analytical scandal while we still grapple with international data transfer policies and AI’s use in decision making. 

Recommended reading

• Tech Professionals: PM Should Make Ethics a Priority at AI Safety Summit
• Survey: UK Workers Embrace Generative AI But Lack Safety Skills
• “Scotland Gets AI”, Innovation Minister Says

But Elyan was more concerned about AI’s impact on a cultural level after noticing a phenomenon he called the “shortcut-driven approach.”

“When I learned programming, we used to write on paper, draw, think about the steps, break things down into components, then go on the computer, and code it, and spend the whole night debugging the code,” he said. 

“Nowadays, the first thing they do is jump on ChatGPT, even if they say they do not, I know that they do,” he said, referring to his students. 

And he’s not sure if this is a good or bad thing – but it does worry him. 

He’s also concerned about the lack of explainability in AI, saying: “In all AI, there’s a philosophy – ‘I don’t care why it works, but tell me it works’ – that’s a problem. We have to do a lot of work on explainability,” he said. 

But his main concern, for Scotland especially, revolves around education. 

“This is what really keeps me awake at night – you have to continuously upskill and reskill,” he said. 

“As educators, we have to rethink how we do things. In Scotland, we have fantastic young people, but I’m not sure how good our education is to uncover their potential.”

“If I was a policy maker, I would put everything else aside, and my task would be to transform education in Scotland.”