Artificial intelligence is already changing how we learn and how we work. The question is no longer whether we will use it, but whether we understand it well enough to use it intentionally.
As an educator researching AI and design pedagogy, I see this tension every day. Students need the freedom to explore technologies that will shape their professions, but they also need to develop the knowledge and judgment to recognize when those technologies are helping them think and when they are thinking for them.
The same is true in the workplace.
The greatest risk isn’t that we use AI. It’s that we stop questioning it.
Literacy Before Adoption
AI literacy is often reduced to technical proficiency: knowing how to prompt a system, generate an image or use the latest platform.
Literacy needs to go further.
We need to understand where AI gets its information, where bias can enter, how outputs should be verified and what happens when the technology is wrong. We need to consider authorship, privacy, ethics and liability. Most importantly, we need to know when, where, why and whether to use it at all.
In my design studios, students use AI to explore narratives, test ideas and identify gaps. But the output is never the end of the process. They document what they accepted, rejected or transformed and demonstrate where their own thinking took the work next.
That process exposes AI’s limitations. Students encounter bias, superficial responses and answers that sound convincing without necessarily being correct. They see similarities emerge across generated outputs and begin asking why.
This is where I’ve seen an important shift: from using AI to find answers to using it to ask better questions. I think of this as augmented design cognition: using AI to extend the designer’s capacity to explore, question and evaluate possibilities without surrendering the human judgement at the centre of the design process.
That distinction is just as relevant in professional practice.
Efficiency Without Over-Reliance
AI can accelerate analysis, summarize information, generate options and remove repetitive work. That creates enormous opportunity.
But efficiency comes with a more complicated question: what happens when we offload the cognitive processes through which expertise develops?
In education, if AI performs the analysis a student is supposed to learn, we have bypassed part of the learning process.
The workplace has its own version of this problem. Early-career professionals develop judgment by researching, drafting, testing alternatives, making mistakes and learning to recognize patterns. Automating those activities may save time, but some of that apparent inefficiency is also how expertise is built.
Not all friction is wasted effort. Sometimes the friction is where learning happens.
This is why I’ve found that restricting AI at certain stages of the design process can be more valuable than unrestricted access. The intention isn’t to prohibit the technology, rather to ensure independent thinking happens before we introduce a tool capable of doing some of that thinking for us.
Organizations can take a similar approach.
Rather than asking only where AI can be used, we should ask where it adds meaningful value, where human capability still needs to be developed and where professional judgment must remain in control.
From Proficiency to Judgment
The AI platforms we use today will change. The more durable capability is judgment.
Can someone recognize when an answer is superficial? Challenge an assumption? Identify bias? Verify information? Understand the implications of acting on an AI-generated recommendation?
More importantly, can they confidently decide not to use AI when their own expertise or another approach would produce a better outcome?
These aren’t separate challenges for education and industry, they’re part of the same continuum.
Universities need to prepare students to enter professions shaped by intelligent systems. Organizations need to ensure that adopting those systems doesn’t erode the expertise and critical thinking those professions depend on.
Both require experimentation. Both require boundaries. And both require literacy before dependency.
AI can create more space for exploration, interpretation and higher-value thinking. But that only happens if we continue developing human capability alongside it.
The measure of successful AI adoption shouldn’t simply be how much we use it or how much time it saves. It should be whether it helps us become better thinkers.