Artificial Intelligence, a useful tool to support human critical thinking

Artificial Intelligence is a very useful tool for humans, but it cannot replace critical thinking, says Peter Rose of TEKenable

Artificial Intelligence and other new technologies are useful tools, but they are accelerators rather than pure decision-making engines; that is where a critical thinking skill set comes in, says Peter Rose of TEKenable.

‘Data Director’ Donal — an office parable

‘Data Director’ Donal fondly recalls his early days in the IT business and still finds it hard to believe that it was more than 40 years ago.

He remembers his first role; a morning milk run pushing a trolly piled high with green bar computer reports, generated overnight on the bright shiny mainframes of the recently arrived Californian-based chip manufacturer. 

 When plonking them without mercy on the desks of overworked accountants and engineers he quietly agreed with his mother’s twice daily wisdom: ‘Work hard. You’ve been lucky to get a start.’ 

Now, retirement day has dawned for Donal, and just one more meeting remains in his emptied calendar. ‘New Grad Gretchen’ has growing concerns over the gathering turmoil known as ‘Artificial Intelligence’ and has asked his advice on how best to use it to serve her lofty career ambitions. Should she reskill, upskill, side skill or just go and do something completely different with her life?

As he mulls his response, his mind’s eye wanders again to some of the technology transformations he has witnessed since that first milk run, all of them caused turmoil too. Dumb terminals, floppy discs, desktop computers fuelled by Seattle-sourced software. Global ERP installations, wide area networks, internet everything, automation, cloud computing, cyber security and many, many more.

And now, just as he reaches for the handle on the exit door, along comes generative artificial intelligence, arguably the greatest technology disruptor since the invention of the transistor. He has no notion what to tell Gretchen but if he asked Peter Rose for advice, it would go like this: “Tell her not to panic. Everything was new and scary once.”

 Like fictional Donal, real-life Peter Rose has enjoyed a long and interesting career in the digital world and founded the Dublin-based TEKenable in 2002, a specialist in Low Code Platforms, Data, AI, and digital transformation services and solutions. 

Most important skill is critical thinking 

Peter Rose’s mission is to help TEKenable’s clients — large and small, domestic and international — to safely navigate the perilous jungle of complex digital transformations.

Peter Rose, founder and owner of TEKenable, an expert in AI, Data, Rapid Digital Transformation and Low Code platforms.

“The skill set doesn't change with AI,” says Rose, warming to his topic. “The skill set is not about understanding individual technologies or the latest ‘state of the art’ development. The skill set is, has always been, and always will be, critical thinking. The ability to decompose a problem into discrete components which can be solved by the tools that you happen to have in your toolbox that day. And then know that your toolbox changes on a regular basis. All that has happened is that we have gained a new tool, and it’s called Artificial Intelligence.” 

Unsurprisingly, the noisy and increasing presence of Artificial Intelligence in the public consciousness means that Rose is extraordinarily busy at present. He and his team fight on the frontline of a shuddering collision between the ‘dry’ digital world, where trillions of ones and zeros are moved rapidly by semi-conductors and stored in data farms, and the’ wet’ world of human beings where emotions are moved rapidly by relentless uncertainty and stored in restless souls.

His goal is to deliver operational clarity, and hopefully, some personal tranquility, to customers whose media headlines scream that ‘the end of the world is nigh,’ humans are about to be enslaved by machines, massive job losses are inevitable, and the consequent societal upheaval will be tumultuous.

Peter Rose, unsurprisingly, doesn’t see things quite that starkly.

“The fact that it looks and sounds and talks like a human being is new. It has the opportunity to be more human-like than any other system has ever been before and that can lead to false expectations and false understandings,” he expands. “AI is probabilistically generated so it's not the final set of steps that lead to a defined outcome. It generates different answers to the same question each time it's asked. Therefore, you must treat it with a high degree of skepticism because it doesn't have the same degree of reliability as previous computer programmes.” 

A subtle but crucial point. In the ‘olden days’ programmes were designed to generate predictable outcomes; if a computer came up with a wrong answer, then it simply had a bug, and bugs could always be fixed. In contrast, Artificial intelligence provides outcomes that are neither predictable nor necessarily accurate. New Grad Gretchen and millions like her will need to adjust to this new ambiguity in their future work where certainty and predictability may be replaced by choice and interpretation.

But first things first, Peter. Will she still even have a job in five years from now? 

“It's not where people think it is necessarily at,” assures Rose. “For instance, in the manufacturing space — most of the automation that's conceivable has already been completed and a lot of the impact of technology on manufacturing has already occurred. The biggest impact is likely to be in areas like mid-management where they're doing things like preparing reports, building content, and so on. That’s the kind of thing that AI can certainly do quicker than a human being. It still needs to be critically reviewed but it's a lot easier and faster to review it than it is to write it from scratch.” 

Yet the unforgiving velocity of the impending AI work upheaval still has countless workers feeling as if they are trying to drink water from a firehose and the feeling of apprehension just won’t go away.

 Additionally, AI programmes now consistently beat the planet’s cleverest science and mathematics PHD graduates in open competition, which opens a whole new vein of uncertainty. The traditional approach of increasing the altitude of education attainment is no longer a guarantee of career progression or employment longevity.

In the face of such change, what practical steps and granular actions would TEKenable recommend to employees such as Gretchen in order to maximise future career opportunities in these new workplaces?

“She should be looking at the technologies she's going to bring to her team or her function,” maintains Rose. “She needs to be familiar with the markets for automation. Gretchen doesn't need to be an AI expert by any means because the AI she is going to be using will be commoditised. The most complex AI technologies will be packaged into a neat model, so she doesn’t have to worry about the level of complexity. She needs to have familiarity with what is possible, what she needs to look out for. There is a need for somebody to supervise that level of control and governance.” 

Next, he considers a constructed example of an employee say, in a traditional credit management function. How will that new AI tool in the toolbox impact day to day work of collecting cash?

“For critical decisions like that, that have wide-ranging impacts, there needs to be a human in the loop and that's the company's opportunity. To understand far beyond the ability of the machine, to understand the situation that the customer finds themselves. That's where a critical thinking skill set comes in. These tools are accelerators rather than pure decision-making engines,” he continues. 

“Because of the probabilistic nature you can't put AI in charge, it can't have an autopilot mode. Dealing with important customers’ needs a human touch and not just a set of rules. Governance and justification for decisions is not going away so you're always going to need a human. You cannot allow an AI service to simply put a customer on credit hold because it has unexpected consequences for that customer, and it could incur potential liabilities for a company that does this.”

 That company is led by a board of directors and senior executives, a cohort often difficult to convince that true change is afoot. The major computing paradigm shifts that both Donal and Peter have known down the years have all been disruptive and expensive and those at the top tables gravely dislike both expense and disruption. The sheer volume of ‘unknown unknowns’ surrounding AI has many CEOs frozen like rabbits in front of oncoming headlights.

When Rose meets with leadership, he first tries to persuade them to avoid ‘error number one’ and view AI as merely a technical problem for delegation to the IT department. He urges them to embrace the AI challenge mainly as a challenge of organisation change, and culture redesign.

“First you need an AI strategy for what is your company going to do in the short and medium term. The long term is far too unpredictable,” he says. “The second thing is education, giving your workforce the skill set to work with these tools so that when they come into your organisation, so that employees understand your intention is to support and assist them, not replace them. 

"Companies are not generally interested in replacing their experienced staff. But you need to level the playing field at Board level and among the senior team as a priority. Define the risks, address the compliance and governance impact, then move on to the efficiency analysis, then change management and education.” 

Despite his dry analysis of the ongoing data revolution, Peter Rose, lives in the wet world too. He has family, friends, employees, customers, and an ongoing stake in a progressive and functioning civic society. So, when his head hits the pillow after yet another busy day, what is the one thing about AI that keeps him awake at night? What is the sum of all his fears?

“I think that I’m already seeing it,” he replies with a chuckle. “It's the cyber security side of things. The dynamic until the very recent past is that human beings were creating ‘malware and phishing’ that were being detected somewhat by software but mostly we relied on the smarts of the person behind the machine. What is now happening very, very quickly is that the machines are finding vulnerability and they are now able to chain vulnerabilities together. 

"The machines are developing ever more sophisticated attacks that humans haven't even thought of, and they move at a pace that is too quick for the human defenders to actually recognise. The AI defenders are in an arms race with the attackers that is already beyond the ability of human beings to recognise. Those with the greatest resources who can train the biggest models with the largest amount of content will win. It's as simple as that.” 

Meanwhile, back to Data Director Donal 

On his very last day in the IT business, Data Director Donal sits alone in his office booking airline seats to the South of Spain and grimaces at the cost of bringing his golf clubs. As he hits the ‘buy’ button he hears a tap on the door and in walks Gretchen. He invites her to sit down, apologises that he only has a few minutes before people will begin to assemble for kind speeches and warm farewells. He notices how fidgety and nervous she seems.

“The only advice I can give you after forty years in this game,” he soothes, “is this. Don’t panic. Everything was new and scary once.”