Only if the system is designed to keep the operator capable. Guidance should help people see what matters and still leave them able to question, override and run the facility without it when conditions change.
Critical facilities are adding layers of machine guidance: alarm prioritization, predictive maintenance, automated load transfer, anomaly detection, and increasingly AI models that recommend actions. Each can make normal operations safer and cheaper. The risk is quieter. Over time, the people who are supposed to supervise the system can lose the ability to do so.
What the research on automation shows
Human factors research has documented this for decades. Parasuraman and Riley's 1997 review in Human Factors described how people misuse automation by over-relying on it and disuse it by ignoring it, with automation bias as a recurring pattern: accepting a system's output even when other evidence contradicts it. Studies of cognitive offloading find the same trade in everyday tools. Offloading makes tasks more efficient and can weaken the skill being offloaded. Dahmani and Bohbot reported in 2020 that habitual GPS use was associated with poorer spatial memory during self-guided navigation.
Anthromekagogy names the facility version of these patterns as failure modes. Automation drift is the gradual removal of human agency through efficiency gains. The skill atrophy loop runs from less practice to weaker competence, to more reliance and further decline.
Design principles that keep people capable
The framework proposes design constraints that translate directly into facility requirements:
- Cognitive integrity. Systems should preserve or strengthen the competencies operators need. In a facility, that means scheduled manual operation of critical equipment, not only automatic tests.
- Calibrated friction. Some effort aids learning and retention, a finding associated with Robert Bjork's work on desirable difficulties. A confirmation step before a consequential automatic action is friction worth keeping.
- Feedback transparency. Operators should know what the system is optimizing, how uncertain it is, and how it fails.
- Reversibility. People must retain the ability to run the facility without the system. When Ukrainian utilities were attacked in 2015, crews restored power by operating substations by hand, which was only possible because they still could.
When there is no time to decide
Some decisions happen too fast for any operator. Protective relays, automatic transfer switches and fire suppression act in milliseconds or seconds. Human judgment still governs them, but it acts at design time, when someone sets what the equipment will do in a situation nobody will be present for. Those settings deserve the same review as any other decision the organization delegates, and the equipment should record what it did so the decision can be examined afterward.
Agency depends on the room
Operators can only exercise judgment if the facility gives them time, information and a safe place to act. An intelligent facility that floods people with alerts while leaving its control room physically exposed does not expand their agency. Physical protection, trusted access and records that survive the event are preconditions for human judgment, not separate topics.
For every new layer of automation, ask what the operator can still do without it, and test it. If the answer shrinks each year, the facility is trading capability for convenience.
Sources
- Parasuraman, R. and Riley, V. (1997). Humans and Automation: Use, Misuse, Disuse, Abuse. Human Factors 39(2).
- Risko, E. F. and Gilbert, S. J. (2016). Cognitive Offloading. Trends in Cognitive Sciences 20(9).
- Dahmani, L. and Bohbot, V. D. (2020). Habitual use of GPS negatively impacts spatial memory during self-guided navigation. Scientific Reports 10.
- Bjork, R. A. (1994). Memory and metamemory considerations in the training of human beings. In Metcalfe and Shimamura (eds.), Metacognition. MIT Press.
- E-ISAC and SANS, Analysis of the Cyber Attack on the Ukrainian Power Grid (2016)
- Kristof, R. (2026). Anthromekagogy: A Research Framework for Human-System Co-Evolution. Working paper.
Cite this article: Authentic Intelligence Research, “Can Machines Guide Operators Without Replacing Their Judgment?” Authentic Intelligence, September 30, 2026, https://authenticint.com/articles/guidance-without-surrendering-judgment.html.