I started my search for LRNT 523 Assignment 1 thinking I was looking for an educational technology researcher. That seemed straightforward enough.

It wasn’t.

The problem was not that I could not find anyone. The problem was that I could find everyone.

My interests currently sit somewhere between educational technology, artificial intelligence, extended reality, firefighter training, motor learning, perceptual expertise, and the split-second read-and-react skills I have spent years observing in both firefighters and athletes. Pick any one of those subjects and there is enough research to keep me occupied for years. Start looking at where they overlap and suddenly the map gets considerably bigger.

This presents an interesting problem for the modern learner. We routinely celebrate the fact that almost the entirety of human knowledge is now available at our fingertips, but we do not talk nearly as much about how we are supposed to get around once we arrive there. Having access to the world’s largest library is wonderful. Wandering through it one bookshelf at a time is another matter.

At some point, the size of the domain changes the usefulness of the tools we use to navigate it.

If the world really is at our fingertips, perhaps taking the bicycle is no longer a virtue.

This is where my use of AI in the MALAT program has become interesting. I am increasingly less interested in having AI answer my questions than I am in using it to discover which questions I should be asking. In that sense, AI has become less like an author and more like a vehicle. I still have to decide where I am going, recognize when I have taken a wrong turn, get out occasionally and look around, and ultimately decide whether the destination was worth visiting. I can simply cover considerably more territory along the way.

Csikszentmihalyi’s (1996) systems model of creativity provides an interesting way of understanding why that matters. Creativity, in his model, emerges through the interaction of the individual, the domain of existing knowledge, and the field that evaluates contributions to that domain. Building on this framework, Csikszentmihalyi and Gruner (2018) extended the systems model to account for artificial intelligence, and Atkinson and Barker (2023) take up that extension to examine how AI can mediate relationships among the creator, the domain, and the field.

That leaves me wondering whether we are asking the wrong question about AI in education.

Perhaps the important question is not simply whether students should be allowed to use the faster vehicle.

Perhaps we should be asking where they went with it.

Friction Is Not Necessarily the Problem

The vehicle metaphor only gets me so far. Going faster is useful, but I do not think the goal of using AI in education should simply be to make the journey easier. If anything, I want the intellectual part of the journey to become harder.

Maybe flight provides a better metaphor.

An aircraft is designed to reduce unnecessary aerodynamic drag, but eliminating its interaction with the air altogether would defeat the purpose. The same air resisting the aircraft is also necessary to produce lift. The aerodynamic shape does not eliminate the forces acting upon the aircraft; it manages those forces so they can be used productively.

I wonder if AI can serve a similar purpose in learning.

There is cognitive friction in academic work that contributes relatively little to the learning I am trying to accomplish. I can spend time checking whether I have italicized the correct part of an APA reference, reorganizing notes from twenty articles, searching repeatedly for terminology I do not yet know, or trying to remember where I encountered a particular concept. Some of that work has value, but some of it simply consumes limited cognitive resources.

Cognitive load theory provides a useful framework here. Sweller (1988) argued that working memory has limited capacity and that instructional activities can consume cognitive resources that might otherwise contribute to learning. Subsequent developments in cognitive load theory have continued to examine how instructional design can manage unnecessary demands while preserving cognitive resources for the complexity inherent in learning itself (Sweller et al., 2019).

AI introduces an interesting possibility into that discussion.

If AI can help reduce some of that unnecessary drag, the objective should not be to create an intellectually frictionless experience.

It should be to redirect the friction.

AI can become the aerodynamic shape that allows the learner to cut through some of the unnecessary drag while preserving—and perhaps deliberately increasing—the forces that generate intellectual lift.

The difficult questions should remain difficult.

I still need to understand the theories. I still need to decide whether the evidence supports an argument. I still need to recognize contradictions, challenge assumptions, compare competing explanations, and decide whether I am making a legitimate connection or simply one that sounds interesting. Those are precisely the places where I want cognitive friction.

Removing those difficulties would not help me soar higher. It would remove the air beneath the wings.

Covering More Ground—and Finding More Questions

My search for LRNT 523 Assignment 1 has become a practical example of this.

The assignment initially appeared relatively narrow: identify an individual outside the course material whose work has made an important contribution to educational technology and examine that contribution. My difficulty was not finding someone who met those requirements. It was deciding which intellectual trail was worth following.

AI allowed me to search with a width and breadth that would have been difficult to achieve through conventional searching alone. I could begin with Seymour Papert and constructionism, move toward Rob Gray’s research involving perceptual-motor expertise and virtual environments, encounter Keith Davids and ecological dynamics, and then consider Joan Vickers’ work on visual attention and the Quiet Eye (Vickers, 2007, 2009). Gray’s (2017) research, for example, examined whether perceptual-motor training undertaken in a virtual baseball environment transferred into real batting performance.

These researchers were not necessarily answering my question. In many cases, they were helping me discover what my question actually was.

That distinction is important.

One question about XR training produced another about proprioception. That raised questions about perception-action coupling. Those led toward representative learning design and ecological dynamics (Chow et al., 2011; Davids et al., 2006). From there came another question: if an XR environment reproduces the visual stimulus but requires the learner to respond through a controller rather than with the movement required in the real environment, what exactly has been simulated?

That matters to me because my interests extend beyond educational technology itself. I am interested in whether XR can reproduce the embodied read-and-react demands found in environments such as volleyball defence or firefighting, where recognizing the correct response is only part of successful performance. The learner must perceive, decide, move, and physically execute the response within the constraints of the environment.

I did not begin the assignment knowing all the terminology required to search for that idea.

That may be one of AI’s most useful functions in research:

It can help me search for things I do not yet know how to search for.

At the same time, there is an obvious danger. Every interesting paper creates another trail to follow, and AI makes following those trails remarkably easy. Curiosity can quickly become wandering.

Ironically, the same tool that broadened my search also helped constrain it.

I could repeatedly return to the actual parameters of Assignment 1 and ask whether the researcher I was investigating still satisfied them. Is this person genuinely connected to educational technology? Is the individual an appropriate subject for the assignment? Can I identify a meaningful contribution? Is there enough scholarship to support the choice? Am I investigating the researcher because they fit the assignment, or simply because I find their work interesting?

AI therefore acted simultaneously as an accelerator and a guardrail.

It allowed me to explore farther outward while repeatedly bringing me back to the road I was supposed to be travelling.

Without exploration, academic research risks becoming an exercise in finding enough sources to support an idea I already had. Without boundaries, exploration can become endless intellectual tourism. The assignment provides the destination and constraints. AI allows me to explore considerably more of the surrounding landscape before deciding which route is worth taking.

Interestingly, the result has not been less curiosity.

It has been considerably more.

Raising the Cognitive Bar

This is why I am not convinced that the appropriate educational response to AI is simply to preserve the amount of effort traditionally required to produce an academic artifact.

We should be more interested in preserving—and increasing—the amount of thinking required to produce it.

If AI helps a learner organize references, locate adjacent scholarship, identify terminology, compare theoretical positions, or check whether a paper complies with APA conventions, some cognitive capacity may become available for other work.

We can reasonably ask what the learner did with it.

Did the learner challenge the literature or merely summarize it? Did they deliberately seek competing perspectives? Did they discover a researcher who caused them to reconsider their original position? Can they explain why one theoretical framework fits their inquiry better than another? Did they identify contradictions? Did they make a defensible connection between domains that are not normally considered together? Did their questions become better as their understanding of the domain increased?

Most importantly:

Did their use of AI increase or decrease the amount of meaningful cognitive work they performed?

That may require professors to expect more rather than less from AI-supported students.

If I can reach Papert, Gray, Davids, and Vickers considerably faster than I could have only five years ago, merely locating and summarizing those researchers should arguably carry less intellectual weight. The expectation can shift toward comparing them, challenging them, identifying the boundaries of their theories, and determining whether their ideas can survive application somewhere they were never intended to go.

AI lowers the cost of reaching those questions.

It should not lower the expectation that I answer them.

In aviation terms, reducing drag is not particularly useful if I use the resulting efficiency only to fly the same altitude and the same distance.

The opportunity is to climb.

We Have Been Wrong About Flying Before

There is some historical irony in using flight as the metaphor.

French military theorist Ferdinand Foch is widely reported to have dismissed the military usefulness of the airplane in its infancy. The exact quotation attributed to him should be treated cautiously because its primary provenance is uncertain, but the story has endured for an obvious reason: history was spectacularly unkind to the sentiment.

The mistake is more interesting than the quotation.

Foch would have been evaluating an immature technology according to what the aircraft in front of him could presently accomplish. At the time, that might even have seemed reasonable. Yet within a few years aircraft were being used for reconnaissance, artillery observation, and aerial combat. The important development was not simply that airplanes became faster or more powerful. Humans discovered entirely new things to do once flight became possible.

There may be a lesson here for how we discuss AI and creativity.

Atkinson and Barker (2023) raise legitimate concerns about AI potentially narrowing the diversity of creative inspiration even while acknowledging its potential as a platform for novel ideas. Those concerns deserve attention. AI can homogenize. It can hallucinate. It can encourage intellectual laziness. It can reproduce what already exists rather than challenge it.

But declaring that AI therefore represents the end or diminishment of human creativity risks making another mistake: judging the eventual creative consequences of a new technology according to the limitations of its earliest forms.

The interesting question may not be whether AI itself is creative.

Nor is it simply whether AI can eventually replace human creativity.

That may be as shortsighted as Foch’s observation.

The more interesting question should be:

“What happens to human creativity once we master flight?”

When the Field Becomes the Bottleneck

There is another problem with Foch’s observation that fits particularly well with Csikszentmihalyi’s model.

Foch was not an uninformed observer standing outside the system. He was an expert. He represented precisely the kind of established field that Csikszentmihalyi argues determines which new ideas are worthy of entering a domain.

That matters because the field serves an essential purpose. Expertise provides quality control. Not every new idea is a good idea, and novelty by itself is certainly not evidence of progress. The field evaluates new contributions against the accumulated knowledge of the domain and determines which ones deserve further attention.

But what happens when the field itself becomes the bottleneck?

I have spent most of my career in a profession that jokes about having “100 years of tradition, unimpeded by progress.” The expression has circulated throughout the fire service for decades precisely because firefighters recognize enough uncomfortable truth in it. Fire-service writers have used the phrase when discussing resistance to changes in technology, training, tactics, and safety (Bashoor, 2018; Pikor, 2016).

There is another fire-service expression that complicates that joke considerably:

Our policies and procedures are written in blood.

That one is not nearly as funny.

The expression reflects an important reality of emergency services. Many of the rules, standards, operating procedures, and safety practices firefighters inherit exist because somebody was killed or seriously injured before us. Sheridan (2011) describes fire-service rules, policies, and standard operating procedures as being “written in blood” because most arose from a member being killed or seriously injured as a result of some prior action, and he singles out the origin of standpipe hose-length policy following a firefighter’s death at a wind-driven kitchen fire as a direct example. The history of live-fire training provides another stark example: the deaths of firefighters William Duran and Scott Smith during a 1982 training fire in Boulder, Colorado, helped drive the development of NFPA 1403 and its requirements for live-fire training (Milan, 2011).

That history changes the way I think about resistance to change.

Tradition in the fire service is not inherently a problem. Neither is conservatism within an academic field. Sometimes the field is resistant because it remembers something the newcomer does not.

There is wisdom in that.

When firefighters tell a new recruit, “We don’t do that,” there may be a very good reason buried somewhere in the history of the profession. Someone may already have tried it. Someone may have been seriously injured doing it. Someone may have died discovering why it was a bad idea.

The field is supposed to remember those things.

This is why simply demanding that established fields “embrace innovation” is too easy. A mature domain contains accumulated knowledge, and some of that knowledge was acquired at extraordinary cost. Progress that ignores that history is not necessarily progress.

But neither is tradition that refuses to reconsider it.

The critical distinction may be between remembering why a rule exists and assuming that the conditions that produced the rule can never change.

“This is what experience has taught us” and “this is how we have always done it” are not the same argument.

The first uses the domain to evaluate a new idea.

The second uses the domain to prevent the new idea from being evaluated at all.

That is where the field can become the bottleneck.

Firefighting provides numerous examples of practices changing as new evidence altered our understanding of the environment. It’s much more than just “Put wet stuff on the red stuff, kid.” Research into modern fire dynamics, flow paths, personal protective equipment, contamination, and firefighter health has challenged practices that were once accepted as part of the job. Fire-service leaders have similarly argued that respecting tradition cannot mean living permanently within it, and that learning from the past is only useful if it produces forward progress rather than a permanent audit of yesterday’s incidents (Bashoor, 2018). Pikor (2016) makes a related point from a technology-adoption angle: fire departments now face so many available tools that the old complaint about tradition unimpeded by progress has, in a sense, inverted—the risk today is not too little technological change but too much of it adopted without a clear sense of what a department actually needs.

The responsibility of the field, then, is more complicated than either protecting tradition or embracing disruption.

The field must be conservative enough to remember why the rules exist and creative enough to recognize when the conditions that created those rules have changed.

That principle transfers remarkably well to education and AI.

There is a legitimate role for professors, institutions, researchers, and academic disciplines—the field—to challenge AI-supported scholarship. Students should be expected to verify sources, recognize hallucinations, understand the theories they invoke, defend their conclusions, and demonstrate that AI has not replaced the intellectual work they were expected to perform.

Some academic practices are “written in blood” too, although thankfully not literally. Requirements surrounding evidence, attribution, peer review, transparency, and methodological rigour exist because scholarship has learned repeatedly what happens when claims are not subjected to scrutiny.

Those protections should not disappear because a new tool has arrived.

But the field should also be willing to ask which academic practices protect intellectual rigour and which simply preserve the mechanics of how intellectual work used to be performed.

If the individual suddenly has a tool that permits much broader and faster access to the domain, then the field may need to reconsider some assumptions about what constitutes rigorous academic work. The traditional difficulty of locating information, organizing references, formatting a paper, or discovering researchers in adjacent disciplines may once have been inseparable from scholarship because there was no practical way to separate them.

AI increasingly makes that separation possible.

Preserving those difficulties simply because previous generations had to endure them risks confusing tradition with learning.

“100 years of tradition, unimpeded by progress” describes what happens when a field becomes so effective at protecting its domain that it begins protecting the domain from change.

Csikszentmihalyi’s field therefore has a difficult responsibility in the age of AI. It must remain skeptical enough to prevent every technologically assisted idea from being mistaken for creativity while remaining open enough to recognize when the tools available to the individual have fundamentally changed what creative participation in the domain can look like.

Perhaps the field should evaluate AI-supported work less by asking whether the learner travelled the traditional route and more by examining where the learner arrived, how they got there, and whether they can defend the journey.

The aircraft has changed.

It would be unfortunate if the field insisted that everyone continue riding bicycles.

Returning to the Creative System

This brings me back to Csikszentmihalyi.

If AI gives the individual greater access to the domain, then access alone becomes a less meaningful demonstration of scholarship. The learner’s contribution increasingly lies in what they do with that access: which connections they recognize, which assumptions they challenge, which ideas they combine, and ultimately what they contribute back for evaluation by the field.

AI therefore does not eliminate the field.

If anything, the field may become more important.

A language model can help me discover Gray, Davids, or Vickers. It can help me compare terminology and identify apparent relationships among their work. It cannot simply declare that my proposed connection between ecological dynamics, XR, volleyball defence, and firefighter decision making represents a meaningful contribution to educational technology. That proposition still has to survive scrutiny from professors, researchers, practitioners, peers, and eventually the larger scholarly community.

The professor therefore remains an essential part of Csikszentmihalyi’s creative system—not simply as an evaluator of the finished artifact, but as a member of the field capable of judging whether the learner has genuinely engaged with the domain.

There is also an interesting paradox here. As AI becomes increasingly capable of navigating and reproducing information already contained within a domain, distinctly human contributions may become more valuable rather than less.

My observations of firefighters learning physical skills do not originate in an academic database. Neither do years of watching athletes learn to read an attacker, recognizing similarities between volleyball defence and fireground decision making, or questioning whether putting someone in a headset and giving them controllers actually reproduces embodied expertise.

Those observations originate with the individual.

AI can help connect those experiences to existing scholarship. It can show me that another discipline has a name for something I have observed. It can expose weaknesses in my assumptions and point toward researchers who disagree with one another. The field can then evaluate whether the resulting idea contributes anything worthwhile.

The creative system therefore remains intact:

Individual experience → AI-assisted exploration of the domain → better questions and novel connections → field evaluation → potential contribution to the domain.

Perhaps, then, the arrival of generative AI should raise rather than lower our expectations of students.

If some of the mechanical drag surrounding scholarship can be reduced, we have an opportunity to deliberately preserve the cognitive forces that produce lift. The goal should not be frictionless learning.

It should be deciding which friction is worth keeping.

The academic question surrounding AI should therefore not simply be, “Did the student use AI?”

We should be asking what happened because they used it.

Did AI allow the learner to avoid intellectual work, or did it allow them to reach intellectual territory they would otherwise never have known existed? Did it simply provide answers, or did it expose assumptions and generate better questions? Did it narrow the learner’s thinking toward a convenient response, or allow them to explore a larger portion of the domain?

If AI gives students unprecedented access to the domain, then perhaps the appropriate response is not to lower the bar or pretend that access does not exist.

Perhaps we raise the cognitive bar.

After all, there is little point in building a better aircraft simply to taxi around the runway.

Now that we can fly, how far—and how high—are we prepared to go?

References

Atkinson, P., & Barker, R. (2023). AI and the social construction of creativity. Convergence: The International Journal of Research into New Media Technologies, 29(4), 1054–1069. https://doi.org/10.1177/13548565231187730

Bashoor, M. S. (2018, April). Chief concerns: Progress doesn’t happen yesterday. Firehouse. https://www.firehouse.com/leadership/article/12397714/learning-from-past-fire-incidents-chief-concerns-marc-bashoor

Chow, J. Y., Davids, K., Hristovski, R., Araújo, D., & Passos, P. (2011). Nonlinear pedagogy: Learning design for self-organizing neurobiological systems. New Ideas in Psychology, 29(2), 189–200. https://doi.org/10.1016/j.newideapsych.2010.10.001

Csikszentmihalyi, M. (1996). Creativity: Flow and the psychology of discovery and invention. HarperCollins.

Csikszentmihalyi, M., & Gruner, D. T. (2018). Engineering creativity in an age of artificial intelligence. In I. Lebuda & V. P. Glăveanu (Eds.), The Palgrave handbook of social creativity research (pp. 447–462). Palgrave Macmillan.

Davids, K., Button, C., Araújo, D., Renshaw, I., & Hristovski, R. (2006). Movement models from sports provide representative task constraints for studying adaptive behavior in human movement systems. Adaptive Behavior, 14(1), 73–95. https://doi.org/10.1177/105971230601400103

Gray, R. (2017). Transfer of training from virtual to real baseball batting. Frontiers in Psychology, 8, Article 2183. https://doi.org/10.3389/fpsyg.2017.02183

Milan, K. (2011, March 7). Live-fire training in acquired structures. Firefighter Nation. https://www.firefighternation.com/fire-leadership/live-fire-training-in-acquired-structures-2/

Pikor, J. (2016, April). Tools & technologies: When worse is better. Firehouse. https://www.firehouse.com/home/article/12171327/tools-technologies-when-worse-is-better

Sheridan, D. P. (2011, August). Qualities of effective incident commanders. Fire Engineering, 164(8). https://www.fireengineering.com/firefighting/qualities-of-effective-incident-commanders/

Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285. https://doi.org/10.1207/s15516709cog1202_4

Sweller, J., van Merriënboer, J. J. G., & Paas, F. (2019). Cognitive architecture and instructional design: 20 years later. Educational Psychology Review, 31(2), 261–292. https://doi.org/10.1007/s10648-019-09465-5

Vickers, J. N. (2007). Perception, cognition, and decision training: The quiet eye in action. Human Kinetics.

Vickers, J. N. (2009). Advances in coupling perception and action: The quiet eye as a bidirectional link between gaze, attention, and action. Progress in Brain Research, 174, 279–288. https://doi.org/10.1016/S0079-6123(09)01322-3

AI Usage Disclosure

AI served as the vehicle described above—faster access to Papert, Gray, Davids, and Vickers, not an author. Every citation was independently verified against the original source afterward, since faster access is no substitute for actually checking the map. The remaining errors and unjustifiably confident metaphors are mine alone; the field is welcome to evaluate accordingly.