Long-term reliance on overlay cues weakens active scanning of the real environment

What it is

Cross on arrows every day, and the head-turn to look both ways happens less and less. The day the headset is off and the arrow is gone, the hole is still in the person. Scanning skill atrophy is the dulling of active search through real structure after environment confirmation has been outsourced to overlays for a long time — not what this frame hid, but that they no longer go looking.

Acute occlusion hurts this time. This hurts the person who is still there after the device comes off.

Why it happens

Skill follows use. Looking both ways before a crossing, checking treads before a stair, checking a threshold before a door, are automated active samples. Overlays turn “confirmation done” into an exogenous signal (green arrow, cleared). Sampling is replaced by that signal. Replacement looks efficient while the overlay is there; once the signal is missing, late, or wrong, the old program is not immediately available. People wait for the next cue, or gamble on a residual scan. Automation complacency and out-of-the-loop in aviation and driving are the same structure: the supervisor no longer spot-checks the layer automation took.

Atrophy needs repetition, not one session. A few days of demo will not show it; weeks and months of a daily route compress scan amplitude, pauses, head-turn range. Worse, people may not feel themselves receding — rated ability can stay high while eye movements have already narrowed.

Studying it

A longitudinal contrast: the same commute or shop-floor route, one group on navigation/warning overlays for weeks, one group without. Stage tests of scanning without overlays (head-turn amplitude, left–right confirm, tread fixations), and misses when the arrow is removed or pointed wrong. Eye and head tracking, plus “do you think you looked,” aimed at the subjective–objective split.

Time: weeks, daily exposure, whether the route is fixed. Dependence: whether the overlay emits an “already confirmed” class of signal. Observations: scan breadth without overlay, correction time after being misled, complacency scales, actual misses.

Turning the overlay off once in a lab measures acute adaptation, not atrophy. Cross weeks, and include a no-device test day.

Where it stops holding

Occasional use (a visit, one repair) does not form a substitute; the product need not change for atrophy. Scanning on a professional monitoring post is a job skill; overlays designed as spot-check aids rather than “already confirmed” atrophy less. Children and novice drivers are still building the scan; overlay substitution interferes with acquisition, more than in practised adults. The hole after doffing appears in real traffic; a lab corridor will not show the cost. Making every confirm a required human tap reduces complacency, and turns every crossing into an operation; weigh it against completion-orientation. It cannot be the only countermeasure.

Applying it

• Safety-confirm cues (you may cross, tread OK) must not be “done” badges that replace a scan. Make them a reminder to look, and require a head turn or a dwell toward the real structure before lighting a pass.
• On a fixed route, insert periodic no-overlay or weak-overlay days so the scanning program is re-enabled. Do not make navigation an unclosable daily prosthesis.
• Wrong cues (a late beat, or a missed point in a training mode) test whether people will scan for themselves. Do not only read hit rate in developer logs.
• How to check: after weeks on a navigation overlay, walk the same crossing one day without the device. If left–right head turns and pauses are clearly below a never-overlay baseline, and a vehicle is missed once, atrophy is already on the road. Still feeling “I was looking” that day is exactly the split.

Related

• Same group：N5.13.1 An overlay hiding a dynamic hazard in the real environment is the highest-grade risk · N5.13.2 While moving, overlay density should drop to leave attention for the environment · N5.13.3 On a performance crash or see-through failure, fallback must prioritize environment visibility · N5.13.4 Stairs, crossings and other safety-critical zones need a rule that clears or fades overlays
• Nearby：N5.07 Overlay Density · N5.12 Overlay Density in Real-world Context
• Search terms：scanning skill atrophy · automation complacency · out-of-the-loop