Rethinking Human Error: What Recent Cognitive Psychology Means for Design
Originally published on Substack
August 10, 2026 · 8 min read
Human Error
For 40 years, UX writing about "human error" has leaned on one framework: Don Norman and James Reason's distinction between slips, lapses, and mistakes (Reason, 1990; Norman, 1984). It's a good framework — durable enough that current HCI research still uses it as shared vocabulary. But treating it as the whole story means missing a decade of research that's more specific, more testable, and in several cases directly actionable for interface design. This article walks through what's changed.
The framework, briefly
Reason's taxonomy splits errors by where they go wrong:
Mistakes — the plan itself was wrong. You intended to do the wrong thing.
Slips — the plan was right, but execution failed at the motor/sensory level (fat-fingering a button, clicking the wrong row).
Lapses — the plan was right, but execution failed in memory (you meant to do it, then forgot).
This still holds up as a root-cause model, and recent HCI papers (e.g. a 2025 review of error-tolerant interface design) build directly on it rather than replacing it. What's newer is the research on why each category happens and what actually reduces it — which is where this update focuses.
Slips and mistakes: bias isn't just a research problem, it's a design lever
A 2025 scoping review presented at CHI mapped how HCI researchers actually study cognitive bias, and the finding worth sitting with is this: interfaces don't just get affected by user biases, they actively trigger them. The review identifies five distinct ways researchers engage with bias — investigating its effects, mitigating it, observing it, deliberately utilizing it, and quantifying it — which is a more useful frame for a design team than a flat list of biases to "watch out for" (Proceedings of CHI 2025).
This matters because it reframes bias from a UX-research hygiene issue (don't let your own bias distort findings) into a product design question (does your interface's structure push users toward predictable errors, and should it?).
On the prevention side, a January 2026 systematic review pooled 20 peer-reviewed studies on ergonomic UI design and put a number on the payoff: error rates in the reviewed studies dropped by roughly a third to two-thirds once ergonomic principles were applied, with people also finishing tasks 20–60% faster — gains concentrated in settings where the stakes and workload were already high (Designs, 2026). That's a meaningfully large effect for what's often treated as a soft design consideration.
On the recovery side — what happens after an error — the foundational argument here actually goes back to Rouse & Morris's 1987 Automatica paper on error-tolerant interfaces. Their core claim: since you can't realistically prevent every error in a complex system, the design goal should shift toward containing the damage once an error happens. They break this into three areas an interface can support — showing the user what an action is currently doing, showing what it's about to do before it's too late to stop, and having the system itself watch for likely mistakes. That framing still shows up as a reference point in current (2025) human-reliability and interface papers. Worth flagging plainly: the recovery-over-prevention argument is older than a "recent research" framing implies; its staying power in current citations is a sign it's held up well, not that it's new.
One design-relevant wrinkle: research on adaptive, AI-driven interfaces (2024) found that dynamic UI can increase skill-based slips, because it disrupts the habitual, muscle-memory patterns users build up — even as it reduces knowledge-based mistakes by surfacing more context. Adaptive personalization isn't a free win against error; it trades one error type for another.
Lapses: the most underrated category, and the most researched right now
Lapses — forgetting to do something you intended to do — map onto what cognitive psychologists call prospective memory, and it's a genuinely active research area.
Two 2025 studies are worth citing directly. One (Applied Cognitive Psychology, 2025) interviewed adults aged 55–86 and found the lapses that bothered people most weren't personal slip-ups, but ones that affected other people — forgetting to meet a friend, for instance. It also found a disconnect between people's confidence in their memory and their actual performance: underconfident in lab tests, overconfident in daily life. Another (Frontiers in Psychology, 2025) found that people's awareness of their own tendency to forget may itself act as a countermeasure — simply knowing you're prone to lapses changes how you compensate for them.
Neither of these is UX research. But both point toward a design implication: users' self-reported confidence about "I'll remember" is not a reliable signal, and systems that assume it is will under-support exactly the people who most need reminders.
Intention offloading: the strongest practical research thread
This is the part of the literature that translates most directly into product decisions. A research program led by Sam Gilbert at UCL has spent the past several years studying how people choose between relying on internal memory and offloading intentions onto external tools — reminders, alarms, to-do lists (Gilbert et al., 2023, Psychonomic Bulletin & Review). A few findings stand out for design work:
People over-offload, systematically. Gilbert's review found a stable bias toward setting more reminders than would actually be optimal, tied more to how confident someone feels in their memory than to their actual memory ability.
Friction suppresses the exact behavior it's meant to support. A 2024 study found that when setting a reminder requires more interaction effort, people become less likely to offload appropriately as a task gets harder — meaning a clunky reminder-creation flow doesn't just annoy users, it measurably degrades their memory performance for the task (Chiu & Gilbert, 2024, Quarterly Journal of Experimental Psychology).
Framing shifts reminder behavior. Presenting offloading as a "gain" (e.g., "stay on top of it") rather than a "loss" pushes people toward a risk-averse pattern of setting more reminders than needed (2022, Cognitive Research: Principles and Implications).
People offload to save effort, not just to avoid forgetting. Reminder-setting increases even when someone's memory for a task is fine, purely because it reduces cognitive load (Sachdeva & Gilbert, 2020, Consciousness and Cognition).
Put together, this is a research base that turns "reduce lapses" from a vague design goal into specific, testable interface questions: how much friction is in your reminder flow, how you word the prompt to set one, and whether your product should nudge toward more offloading or actively discourage over-reliance on it.
The bigger picture: dual-process theory hasn't been abandoned, it's being refined
Underneath all of this sits dual-process theory — fast, automatic System 1 thinking versus slow, deliberate System 2 thinking (Kahneman, 2011; Evans & Stanovich, 2013). Recent work isn't discarding this model but complicating it usefully. A 2025 Frontiers in Psychology study proposed a "Dual Process Model 2.0," using cognitive-load manipulations to show that intuitive, System-1-style judgments aren't automatically less valid than deliberate ones — intuition can itself be logical under the right conditions. A late-2025 integrative review of decision fatigue also reinforces why error rates climb under pressure and cognitive load in high-stakes fields, citing NASA data attributing roughly 80% of aviation accidents to human decision-making errors under uncertainty.
For design purposes, the practical takeaway is that "the user is in System 1 mode" isn't a sufficient diagnosis on its own — cognitive load, framing, and confidence all shape whether fast thinking produces a good or bad outcome, and each is something an interface can influence directly.
What this means for practice
Treat slip/lapse/mistake as the skeleton, not the whole model — layer in why each occurs and what specifically reduces it.
Audit reminder and notification flows for interaction friction; it has a measured, causal effect on whether people offload appropriately.
Be deliberate about framing in prompts that ask users to set reminders or confirm intentions — framing measurably shifts behavior.
Don't rely on users' self-reported confidence ("I'll remember") as a design signal — it's a weak predictor of actual lapse rates.
If you're introducing adaptive or AI-personalized UI, expect a trade-off: fewer knowledge-based mistakes, but a possible rise in skill-based slips as habitual patterns get disrupted.
For error recovery flows, design for detection, explanation, and correction as distinct stages, not just a single "error state."
Sources
Reason, J. (1990). Human Error. Cambridge University Press.
Norman, D. A. (1981). "Categorization of Action Slips." Psychological Review, 88(1), 1–15. https://doi.org/10.1037/0033-295X.88.1.1
"How Do HCI Researchers Study Cognitive Biases? A Scoping Review." (2025). Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems. https://dl.acm.org/doi/10.1145/3706598.3713450
"The Impact of User Interface and Experience (UI/UX) Design on Visual Ergonomics: A Technical Approach for Reducing Human Error in Industrial Settings." (2026). Designs, 10(1), 8. https://doi.org/10.3390/designs10010008
Rouse, W. B., & Morris, N. M. (1987). "Conceptual design of a human error tolerant interface for complex engineering systems." Automatica, 23(2), 231–235. https://www.sciencedirect.com/science/article/pii/0005109887900975
"A New Era in Human Factors Engineering: A Survey of the Applications and Prospects of Large Multimodal Models." (2024). arXiv. https://arxiv.org/pdf/2405.13426
Nygaard, A., et al. (2025). "Understanding Older Adults' Experience of Prospective Memory Errors and Strategy Use." Applied Cognitive Psychology, 39, e70066. https://onlinelibrary.wiley.com/doi/full/10.1002/acp.70066
(2025). "Prospective and retrospective memory lapses and academic self-efficacy." Frontiers in Psychology. https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1515533/full
Gilbert, S. J., Boldt, A., Sachdeva, C., Scarampi, C., & Tsai, P.-C. (2023). "Outsourcing Memory to External Tools: A Review of 'Intention Offloading.'" Psychonomic Bulletin & Review, 30(1), 60–76. https://link.springer.com/article/10.3758/s13423-022-02139-4
Chiu, G., & Gilbert, S. J. (2024). "Influence of the physical effort of reminder-setting on strategic offloading of delayed intentions." Quarterly Journal of Experimental Psychology. https://journals.sagepub.com/doi/full/10.1177/17470218231199977
"Framing cognitive offloading in terms of gains or losses: achieving a more optimal use of reminders." (2022). Cognitive Research: Principles and Implications. https://link.springer.com/article/10.1186/s41235-022-00416-3
Sachdeva, C., & Gilbert, S. J. (2020). "Excessive use of reminders: Metacognition and effort-minimisation in cognitive offloading." Consciousness and Cognition. https://www.sciencedirect.com/science/article/abs/pii/S1053810020300179
Zucchelli, M. M., et al. (2025). "The Dual Process model: the effect of cognitive load on the ascription of intentionality." Frontiers in Psychology, 16, 1451590. https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1451590/full
"An integrative review on unveiling the causes and effects of decision fatigue to develop a multi-domain conceptual framework." (2025). Frontiers in Cognition. https://www.frontiersin.org/journals/cognition/articles/10.3389/fcogn.2025.1719312/full
Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
Evans, J. St. B. T., & Stanovich, K. E. (2013). "Dual-Process Theories of Higher Cognition: Advancing the Debate." Perspectives on Psychological Science, 8(3), 223–241. https://doi.org/10.1177/1745691612460685
Share
A small favor
Was anything I wrote confusing, outdated, or incorrect? Please let me know! Just write a few words below and I’ll be sure to amend this post with your suggestions.