Should we learn AI in the AI age or just let AI do its thing?
How did we arrive at a point where that is even a question worth asking? Does this mean that the artificial intelligence (AI) revolution has demoted learning technology to an afterthought, or a ‘requirement’ to be relevant in the evolving job market? TechPros Marketing and Communications Lead, Shiba Kurian, led the panel discussion.
If Artificial Intelligence (AI) can do it for you, should you still learn how to do it yourself? TechPros Rotterdam recently posed this question to a group of experts at a panel discussion in the city. The answer was unanimous: “Yes, of course”. It was not a straightforward ‘yes’.
The idea to broach the topic of ‘learning AI’ in the age of AI stemmed from the ubiquitous trend of a rush to sell and learn an arsenal of AI and related tech courses. As a tech education organisation, it made us pause and ask: Why are we offering our courses? Are they really enhancing and accelerating tech skills? Are our learners enrolling out of genuine curiosity and passion for developing a new tech skill, or out of fear of being replaced or of companies sidelining those who do not adopt it fast enough? How are we approaching ‘learning’? It was the leitmotif of the panel discussion.
The flawed approach to learning
The panellists pointed to an unhealthy phenomenon gripping the educational sector: education has shifted towards viewing learning as an outcome (think diploma, specific skill set, credits) rather than a process. This makes AI feel threatening, according to the panellists.
“If a skill is just an end result, and AI can approximate that end result, then it looks like AI has already replaced the human,” said Marvin Kunz, an AI developer with a background in behavioural psychology.
It takes a reductive approach to actual jobs and skills, in essence, dehumanising them. Marvin elucidates his point with the Doorman Fallacy. On paper, a hotel doorman’s job is simple: open and close the door. It can easily be automated with a motion-sensor door. But in reality, the doorman also greets guests, deters unwanted behaviour and visitors, and sets the tone or atmosphere of the hotel – things an automatic door (or AI) cannot replicate. This fallacy, which reduces and evaluates jobs to their narrowest definable output, is at the heart of the “AI will replace humans” narrative. “So every job is far more relational, contextual, and judgment-based,” said Marvin.
AI is currently framed as a historically unprecedented power technology, built on the premise that humanity is flawed and the technology will save us. But that is misleading, Morraya Benhammou, a senior lecturer, educational innovator, and AI didactics specialist at the Hague University of Applied Sciences, challenged. “What the steam engine was to people decades before us is what AI is to us now. Every generation witnesses a technology that feels world-altering at the time. AI is not categorically any different. I see it as a foundational technological knowledge,” according to Morraya.
As an educator, she recognises the contradiction of being part of an education system that pushes students toward a job market that is bloated with fixed requirements. “On the one hand, I am empowering students to do something for their future. On the other hand, I am part of a system that says, ‘there is no job for you because things will be automated’. It fuels insecurities and fear. AI is just a technology that humans created; we should treat it as such,” she said.
The challenge today predates AI, the panellists noted. Historically, students built critical thinking and broad reasoning skills (philosophy, arts, etc.) before specialising. Gradually, learning outcomes have trickled down from some market-driven entity on roughly a five-year cycle. “Now, students are dropped straight into narrow domains such as marketing, finance, and information management, with no foundation in how to think,” said Morraya. That said, domain knowledge still matters, especially with AI in the mix.
“But the real aim should be preparing students to be disruptors, not casualties of disruption, or destructees. That requires moving away from treating education as a series of hoops to jump through.”
According to Joana Kroon, Global Project Manager of Operations at Cambridge Innovation Center CIC Rotterdam, when workplaces assign tasks to an employee, with no instruction manual, some are often confused about how to approach it. While asking the right questions to arrive at a decision is a skill in itself, Joana traces this back to one’s upbringing, not just schooling. Drawing on her own African background, Joana notes she was not encouraged to ask questions as a child. Her curiosity was often dampened with “this is what it is” or “be quiet”. She had to consciously unlearn that, including the assumption that seniority or authority equals knowing the answer, like how AI claims to know it all.
Redefining learning in the age of AI
The urgency to learn AI has two distinct drivers: fear of job loss and curiosity to understand how to leverage AI to improve their work and lives. This dual motivation is not necessarily a bad thing, according to Maaike Wachters, a software engineer and TechPros Rotterdam Chapter Lead. “It is actually good that people are engaging with AI, since it is becoming an integral part of both jobs and everyday life. What matters is how people engage with it. AI should be approached as a tool to improve learning, not simply out of panic,” she cautioned.
Critical thinking is the number one skill for the AI age, according to Maaike. She encouraged using AI as a tool, not a thinking replacement. “AI is bad at thinking for you but very good at processing and summarising large amounts of information. Use AI for that, and reserve judgment and reasoning for the human side. Combine the two well, and you go further,” she said.
Prodding further on the critical point, Maaike observed that the real problem is not AI itself, but rather the tasks we assign to learners and students. For example, if you ask a student to ‘write a paper’, they will turn to ChatGPT to write it, because people are inherently inclined to avoid effort if a shortcut exists,” according to Maaike. Of course, this is not a student-specific problem. The same logic applies to anyone. If a task can be offloaded, it will be.
The fix: Change the task, not just the prohibition. “If instead you ask the student for their opinion and have a live conversation rather than demanding a written paper, they are forced to actually use critical thinking, because there is no shortcut for that,” Maaike pointed out.
Some love learning, and some do not like learning, regardless of AI. If a person approaches learning as ‘I am going to school for the diploma’, and not ‘learn because I want to understand certain things’, then AI only generates a pre-defined output; nothing new is created, and the existing hollow pattern gets reinforced and scaled up. That is the danger of letting AI do its thing. As researcher Kentaro Toyama’s thesis notes, it is the positive human intent that can lend AI or any tech an additive or transformative effect; otherwise, it only amplifies the existing lack of motivation rather than fixing it. Good habits get better; shortcuts get shorter.
The positive human intent, in terms of learning, involves a person approaching a topic with passion, including searching the internet, reading papers, articles, and books, and engaging with AI to explain certain incomprehensible concepts, according to Joana. AI then accelerates the learning. This is essential, especially at a time when AI is premised on the supposed promise that it makes our lives easier and more productive. Despite this lure of tech, we find ourselves busier and more fatigued than ever, building ever more AI systems. That is where a positive intent to learn could ensure AI makes us smarter (read ‘a good shortcut’) rather than more overwhelmed.
For instance, Joana uses AI at work for time-consuming tasks that are not central to the actual value or purpose of the business. She turns to vibe coding to create standard operating procedures (SOPs).
“I do not want to sit for hours on the computer for a standard task that is built on standard formats. Instead, I spend time talking to customers and understanding their needs, because that is the function of my job. Essentially, let AI do what it is best at, and let me do what I am best at. This has helped me positively in my career development,” said Joana.
The discussion on the risk of every watershed technology rendering today’s skills obsolete also steered the conversation to upskilling: How long will the cycle of constant upskilling continue? However, as Morraya argued, micro-credentials and lifelong learning have always existed. In fact, these forms of upskilling allow professionals to bring real problems from their work environments into their learning, making upskilling more contextual and self-directed rather than rigid or generic.
The need for formal upskilling also depends on where you want to work, according to Marvin. Large corporations/institutions, for example, tend to require formal certificates as a baseline requirement. Smaller, dynamic organisations often care more about demonstrable projects and practical ability than credentials.
Besides, many skills are intangible and hard to ‘formally’ teach. Skills are often not learned in institutional settings at all, but through conversation, practice, and real-world exposure. Institutions can still play a role in signalling which skills matter. “A lot of valuable learning happens informally – on the job, through customised training, or through workplace budgets, which can even be used for non-job-related wellbeing, like yoga or gym,” said Morraya.
In essence, before embarking on the next tech course, panellists encouraged one to ponder: what is the end goal of education or learning? Is it turning in the paper, getting credits, and earning a diploma?
So, yes, we should learn AI, but learning should be driven by genuine curiosity, not tied to a specific job outcome or mandated by a career requirement.
