Why the "Weapon Progression" Analogy Falls Short: AI, Learning Psychology, and Intellectual Development in Higher Education

An Instrument Box

(Courtesy of Ms Geok Lee)

 The analogy that students should first master "simple tools" before being allowed to use artificial intelligence (AI) assumes that learning is fundamentally linear and hierarchical. Like progressing from a pistol to an automatic rifle, then to tactical weapons, and eventually to a nuclear weapon, students are expected to demonstrate competence at one level before advancing to the next. The underlying assumption is that greater technological power should only be entrusted to individuals who have first mastered less powerful tools.

Although intuitively appealing, this analogy oversimplifies how learning, expertise, and intellectual development actually occur in contemporary higher education. More importantly, it misunderstands the role that AI plays in human cognition. Weapons amplify destructive capability. AI, by contrast, is a cognitive technology designed to augment human thinking, problem-solving, and decision-making. Treating AI as though it were merely a more powerful "weapon" ignores decades of research in educational psychology, cognitive science, and professional education.

From the perspective of the psychology of learning, knowledge acquisition is rarely a straight, sequential process. Modern theories describe learning as iterative, recursive, and socially mediated rather than rigidly hierarchical. Learners continuously move between acquiring foundational knowledge, applying concepts, receiving feedback, reflecting on errors, revising misconceptions, and reconstructing their understanding. They revisit concepts repeatedly, each encounter deepening and reorganizing their knowledge. Learning is therefore less like climbing a ladder than navigating a complex network of interconnected ideas.

Lev Vygotsky's concept of the Zone of Proximal Development (ZPD) illustrates this particularly well. According to Vygotsky, students often develop their highest levels of understanding when supported by more capable partners or cognitive tools. Traditionally, these supports included lecturers, tutors, peers, textbooks, laboratories, and educational technologies. Today, AI represents another form of cognitive scaffolding. It can explain difficult concepts, generate examples, pose alternative perspectives, provide immediate feedback, identify misconceptions, and encourage reflection. The educational question, therefore, is not whether students should use AI, but how AI should be integrated into learning environments so that it enhances understanding without replacing genuine intellectual effort.

Jerome Bruner similarly argued that effective education depends on scaffolding. Learners are provided with appropriate support while they are developing competence, and this support is gradually withdrawn as they become increasingly independent. Importantly, scaffolding does not require withholding sophisticated tools until learners have achieved complete mastery. Rather, it recognises that carefully designed assistance enables learners to perform beyond what they could accomplish alone. AI can serve precisely this function when thoughtfully integrated into teaching and learning activities.

Constructivist theories reinforce this perspective by emphasising that learners actively construct knowledge through exploration, experimentation, questioning, discussion, and revision. Students frequently encounter ideas that exceed their current understanding. They make mistakes, test hypotheses, and gradually refine their mental models. This productive struggle is not evidence of poor learning, it is one of learning's defining characteristics. Access to sophisticated cognitive tools during this process can deepen understanding, provided students remain intellectually engaged rather than becoming passive recipients of AI-generated answers.

Research on expertise further challenges the assumption that intellectual development follows a fixed sequence. Experts do not simply know more facts than novices. They organise knowledge differently, recognise meaningful patterns more efficiently, and strategically combine internal knowledge with external resources. Professionals routinely consult databases, decision-support systems, simulation software, statistical packages, design platforms, legal research systems, and increasingly AI. Expertise therefore involves knowing when and how to use advanced tools and not merely functioning without them.

Consequently, universities should prepare students to develop sound judgment in the use of AI rather than delaying exposure until some arbitrary threshold of competence has been reached. Graduates entering contemporary workplaces will encounter AI-integrated environments regardless of discipline. Shielding students from these technologies during their education risks producing graduates who are technically underprepared for the realities of professional practice.

History repeatedly demonstrates that powerful technologies do not necessarily diminish learning. Similar concerns accompanied the introduction of calculators, statistical software, search engines, computer-aided design systems, computer algebra systems, and integrated programming environments. Critics feared that automation would weaken fundamental skills. Instead, these technologies shifted educational priorities. As routine computational or procedural tasks became automated, greater emphasis moved toward conceptual understanding, interpretation, evaluation, creativity, systems thinking, and complex decision-making. AI is likely to produce a similar transformation.

The evolution of aviation provides a particularly illuminating analogy. During the early era of commercial aviation, long-haul aircraft typically required a cockpit crew of five, a captain, a first officer, a flight engineer, a navigator, and a radio operator. Each role performed specialised tasks essential to safe flight. Over time, however, advances in avionics, onboard computers, satellite navigation, digital communications, sensors, and automated flight management systems gradually assumed many of these responsibilities. Modern commercial airliners are routinely operated by only two flight crew members, the captain and the first officer.

This transformation did not occur because the responsibilities of navigation, systems management, fuel balancing, communications, or performance monitoring became less important. On the contrary, these functions remain absolutely critical to flight safety. What changed was that computers became better at continuously monitoring, calculating, and managing routine operational tasks. Human expertise consequently shifted from manually performing every task toward supervising automated systems, interpreting information, exercising judgment during abnormal situations, and making complex decisions when automation reaches its limits.

Modern aviation therefore illustrates a broader reality of twenty-first-century work, where computers increasingly manage systems while humans supervise computers. The pilot's role has not disappeared, instead it has evolved. Today's pilots require an even deeper understanding of aircraft systems, automation logic, human factors, risk management, and decision-making than their predecessors. Their expertise lies less in manually calculating every navigation point or fuel transfer and more in understanding when automation should be trusted, when it should be questioned, and when human intervention becomes necessary. This evolution closely mirrors the educational challenge posed by AI. Universities should not train students for a world in which they compete against computers in routine cognitive tasks, they should prepare students to supervise, critique, and collaborate effectively with intelligent systems.

Military technology offers a similar lesson. Contemporary battlefields increasingly employ unmanned aerial vehicles (UAVs) whose missions are controlled remotely by personnel located thousands of kilometres from the battlefield. The individual directing a UAV need not physically occupy the aircraft or even be present in the theatre of operations. Instead, technicians or “warfighters” operate sophisticated sensor systems, communications networks, targeting software, and autonomous flight technologies from secure control stations. Their effectiveness depends not on manually flying the aircraft in the traditional sense, but on interpreting data, making timely decisions, coordinating information, and exercising sound judgment within highly automated systems.

This development demonstrates an important distinction. Modern military effectiveness depends less on manual operation of increasingly complex machinery and more on the ability to manage sophisticated human-machine systems. The expertise required has shifted from direct physical control to cognitive supervision. The same transformation is occurring across medicine, engineering, finance, logistics, manufacturing, law, scientific research, and education. AI is not replacing human intelligence, it is changing the kinds of intellectual work humans are expected to perform.

Intellectual development in higher education similarly encompasses far more than procedural competence. William Perry's model of intellectual and ethical development suggests that university students gradually progress from viewing knowledge as fixed, certain, and authority-driven toward recognising ambiguity, evaluating competing evidence, and making justified judgments under conditions of uncertainty. AI, when critically interrogated rather than accepted uncritically, can actually stimulate these higher-order forms of reasoning. Students may compare AI-generated explanations with scholarly sources, identify factual inaccuracies, evaluate competing interpretations, refine prompts, challenge assumptions, and defend conclusions using evidence. These activities cultivate precisely the kinds of analytical and reflective thinking that higher education seeks to develop.

The weapon progression analogy also fails because it assumes that access to powerful technology necessarily reduces learning. In reality, educational technologies change the distribution of cognitive effort rather than eliminate it. The routine production of information becomes easier, while the evaluation, interpretation, integration, ethical application, and communication of knowledge become more demanding. Students still require strong disciplinary foundations, but those foundations increasingly support higher-order reasoning rather than repetitive procedural execution.

A more appropriate analogy, therefore, is learning to operate within increasingly intelligent technological ecosystems rather than progressing through increasingly destructive weapons. Pilots train extensively using sophisticated flight simulators from the earliest stages of instruction. Medical students interact with advanced imaging technologies, robotic surgical systems, and computer-assisted diagnostic tools throughout their education while simultaneously learning anatomy, physiology, pathology, and clinical reasoning. Engineering students use professional modelling software long before becoming licensed engineers. Scientists rely on computational modelling, data analytics, and AI-assisted discovery while continuing to apply rigorous scientific reasoning. In each case, powerful technologies do not replace foundational knowledge instead they provide authentic contexts in which foundational knowledge is developed, tested, and applied.

This does not imply that unrestricted or uncritical AI use is educationally desirable. Students still need opportunities to develop independent reasoning, disciplinary expertise, foundational conceptual understanding, and the ability to solve problems without technological assistance when appropriate. There remain contexts in which unaided performance is pedagogically valuable, just as pilots continue to practise manual flying despite widespread automation. The objective is not technological dependence, but technological fluency combined with intellectual independence.

These goals are best achieved through intentional instructional design rather than blanket prohibition. Assessments can require students to explain their reasoning, critique AI-generated responses, document and justify their prompting strategies, compare multiple solutions, identify AI limitations, evaluate evidence, and defend conclusions through disciplined argumentation. Such assessments measure genuine understanding rather than simple content production. They encourage students to think with AI without allowing AI to think for them.

Ultimately, the military progression analogy misunderstands both the nature of AI and the psychology of learning. Weapons extend destructive capability whereas AI extends cognitive capability. Learning in the twenty-first century is not a linear march from simple to complex but a dynamic interaction among learners, educators, peers, knowledge resources, and increasingly intelligent technologies. The historical evolution of aviation demonstrates that technological progress does not eliminate expertise but it transforms it. Likewise, the rise of UAVs illustrates that modern competence increasingly involves supervising sophisticated systems rather than manually executing every task. Across professional domains, humans are moving from being operators of technology to supervisors, interpreters, and decision-makers within technologically mediated environments.

Higher education should therefore focus not on delaying students' access to AI but on cultivating the intellectual maturity, ethical judgment, disciplinary expertise, and critical thinking required to use AI wisely, responsibly, and creatively. The central educational challenge is no longer whether students should use AI, but whether universities can prepare graduates who know when to trust AI, when to question it, when to override it, and ultimately, how to remain accountable for the decisions made with its assistance.

This argument does not suggest that AI should replace foundational learning. Rather, it contends that universities should teach students to think with AI without allowing AI to think for them. That distinction is far more consistent with contemporary theories of learning, intellectual development, professional education, and the realities of an increasingly AI-mediated world.

Cheers.

ravivarmmankkanniappan@132920072026 3°3′52′′N 101°35′37′′E

©All Rights Reserved.