A couple of years ago, if you asked an AI LLM to build you a differentiated lesson plan, there was a good chance it would have produced activities sorted by learning style: something visual for your visual learners, something kinesthetic for the students who need to move. The idea has been contested in the research literature for the better part of two decades.
That is a useful reminder of how learning myths actually survive. They persist because they are intuitive, because they are rooted in good intentions, and because each new delivery mechanism gives them somewhere fresh to live. Personalizing instruction to a student's learning style sounds respectful of individual differences. Giving students the freedom to discover concepts independently sounds like an ideal way to foster curiosity and ownership.
But what feels right and what supports evidence-based learning are not always the same thing. Understanding that distinction can help us design instruction around how students actually learn.
Myth 1: Students learn best when instruction matches their learning style
For years, educators have been encouraged to think of students as visual, auditory or kinesthetic learners and tailor instruction accordingly. The research has not supported it. In a review commissioned by Psychological Science in the Public Interest, Pashler, McDaniel, Rohrer and Bjork found that almost no studies had used the experimental design capable of actually testing the matching hypothesis, and among the few that did, the results ran contrary to it.
That doesn’t mean visuals, discussion, movement or hands-on experiences aren’t valuable. The distinction is that we should select instructional approaches based on what students are learning, rather than assigning students to categories based on how we believe they learn best.
A diagram may be effective for explaining the structure of a cell. Hearing pronunciation matters when learning a new language. Manipulatives can make an abstract mathematical relationship more concrete.
What should educators do instead? Choose the modality that best serves the content and learning objective. Rather than asking, “What kind of learner is this student?” ask, “What is the clearest way to help this student understand this concept?”
Myth 2: Students don't need to memorize facts they can look up
We shouldn't spend class time on something a student can google or ask AI, right? If a model can retrieve any fact in seconds, the reasoning goes, class time is better spent on analysis and evaluation.
But thinking is not a content-free skill. Daniel Willingham's formulation is that factual knowledge precedes skill, and that apparent skill deficits are usually knowledge deficits in disguise. You cannot evaluate an argument in a domain whose basic claims are unfamiliar to you. Knowledge held in long-term memory also doesn't compete for working memory; it expands what working memory can do. A student who has automated multiplication facts has capacity available for the structure of a problem. A student who is still calculating has already spent it.
What should educators do instead? Build knowledge deliberately and sequence it, rather than deferring content in favor of generic skills instruction. The higher-order work we want from students rests on a foundation, and that foundation cannot be outsourced to a search bar.
Myth 3: We only use 10% of our brains
Few brain myths have proved as durable as the idea that humans use only 10% of their brains.
The myth implies we are sitting on vast reserves of untapped capacity, waiting for the right technique to unlock them. The reality runs in precisely the opposite direction. Our cognitive resources are sharply limited, particularly attention and working memory. John Sweller's cognitive load theory rests on this constraint: working memory holds only a handful of new elements at once, and everything a student is asked to process that isn't the content is capacity spent on something other than learning.
What should educators do instead? Manage cognitive demands deliberately. Break complex material into manageable steps, connect new concepts to prior knowledge, remove unnecessary distractions and revisit important learning over time.
Myth 4: Students learn best when they discover concepts for themselves
This misconception sits at the intersection of several things educators rightly value: curiosity, autonomy, inquiry and student ownership. These goals matter. But student agency should not be confused with minimal instructional guidance. Kirschner, Sweller and Clark later made the cognitive case for why, arguing that minimal guidance ignores what we know about human cognitive architecture, and that the advantage of guidance recedes only once learners have enough prior knowledge to guide themselves. Students need sufficient knowledge, modeling and instructional guidance before we ask them to independently discover complex concepts.
Imagine asking a novice to solve a complicated problem while simultaneously determining which information matters, selecting a strategy and learning unfamiliar content. Working memory can quickly become overwhelmed.
Explicit instruction offers another path, and it should not be confused with passive lectures. It can include breaking learning into manageable steps, modeling worked examples, thinking aloud, leading guided practice, checking for understanding and gradually releasing responsibility. Retrieval practice and spaced review can then strengthen learning over time.
Inquiry, exploration and problem-solving still have an important place. We simply set students up to benefit from those experiences when we first provide the knowledge and scaffolding, they need to succeed.
Look beyond instruction to the conditions surrounding it
Correcting misconceptions about learning is only part of the work. We also need to examine the environment in which learning occurs. Nobel laureate Herbert Simon once described watching an ant make its way across a beach. Its path may appear extraordinarily complicated, filled with turns and detours. But much of that complexity isn't coming from the ant; it is responding to obstacles in its environment.
I think about students in much the same way. We can set high expectations and ask students to focus or persist. But sometimes the student is already expending enormous effort navigating barriers we've placed in the path of learning. Our responsibility is to ask: Which barriers can we remove?
This could mean providing clearer modeling, breaking a complex task into smaller steps or minimizing distractions competing for students’ attention. It can also mean examining the physical classroom. If students must expend cognitive energy trying to hear and understand their teacher over competing noise, that is energy unavailable for learning. My work with Lightspeed has reinforced my interest in designing learning environments where every student can clearly access instruction.
Before assuming students need to try harder, we should examine whether the environment is making learning unnecessarily difficult.
Apply the same scrutiny to what comes next
These older learning myths also offer a lesson as educators navigate AI. We are once again encountering tools and ideas evolving faster than our understanding of their implications for learning. AI can even perpetuate outdated concepts; ask it to create differentiated lessons, for example, and it may still recommend activities based on “learning styles.” Something being new, popular or technologically sophisticated does not make it instructionally sound.
Before adopting a new strategy or tool, educators can ask:
● What evidence supports it?
● What prior knowledge does it assume?
● What demands does it place on attention and working memory?
● What guidance will students need before working independently?
● What barriers could interfere with learning?
Learning science will continue to evolve, and our practices should evolve with it. The goal isn't to replace one set of educational absolutes with another. It is to examine evidence, question assumptions and adjust our practice as our understanding grows. Ultimately, the question isn't whether an instructional approach sounds engaging, personalized or innovative. It's whether we are creating the conditions students need to learn.
About the Author
Dr. Nathan Lang-Raad is an educator, speaker, author, innovator, an advocate for AI that strengthens human connection, and a VP of business development for Lightspeed.
