21.18 Human Machine Communication Error

A human-machine communication error is any failure in the exchange of information between a human user and a machine system that results in an unintended outcome — a misunderstanding, a failed task, a wrong action, a safety-critical incident, or a degraded interaction experience. Human-machine communication errors are not limited to technical failures of the machine; they include errors arising from the human side of the interaction, from the design of the interface that mediates the interaction, and from the structural properties of the human-machine communication system as a whole. Understanding human-machine communication errors requires analyzing the entire communication loop — what the human intended, what the machine received, what the machine produced, and how the human interpreted the machine's response — to locate where the breakdown occurred and what conditions produced it.

Classification of Human-Machine Communication Errors

Human-machine communication errors can be classified along several dimensions:

By originating party: Some errors originate on the human side — the user issues an incorrect command, misspecifies parameters, selects the wrong action, or misinterprets a correctly presented output. Other errors originate on the machine side — the system processes a correct input incorrectly, produces an inaccurate output, fails to respond, or generates a misleading feedback signal. Many errors are genuinely bilateral — arising from the interaction between human behavior and machine design in ways that are not fully attributable to either party alone.

By detection: Detected errors are failures that the user or system recognizes as errors and attempts to address through recovery and repair. Undetected errors are failures that neither the user nor the system identifies as errors — the user proceeds with a false belief about what the system has done, or the system proceeds with a false interpretation of what the user intended, without any correction being initiated. Undetected errors are typically more consequential than detected ones because they have the opportunity to compound through subsequent actions taken on incorrect premises.

By severity: Communication errors range from minor inconveniences that slightly delay task completion to catastrophic failures with serious real-world consequences. Severity depends on the function of the system, the stakes of the task, the reversibility of the action triggered by the error, and the time available for detection and correction before consequences become irreversible.

Specific Error Types

Several named error types recur across human-machine communication research:

Mode errors occur when a user's action is appropriate for one system mode but the system is in a different mode — the action is interpreted differently than the user intends because the system's current mode is not what the user believes it to be. Mode errors are prevalent in complex systems with multiple operating modes and inadequate mode visibility. The user acts on the wrong model of the system's current state.

Confirmation errors occur when a user confirms an action without reading or understanding the confirmation dialog — clicking through a confirmation prompt that warns of consequences the user has not registered. Confirmation errors are particularly common when confirmation prompts are routine and frequent, causing users to habituate and respond automatically to the prompt format rather than reading its content.

Encoding errors occur when a user intends to specify one value and enters a different one — transposing digits, selecting the wrong item from a list, or entering a value in the wrong field. Encoding errors represent failures at the level of input specification rather than intent, and are particularly likely under time pressure or attention division.

Interpretation errors occur when a user misinterprets a system output — reading a status display incorrectly, misunderstanding the scale of a chart, or parsing an error message as indicating a different condition than it actually reports. Interpretation errors are on the human side of the communication but are often caused by output design that fails to match user interpretive frameworks.

Omission errors occur when a user fails to perform a required action — missing a step in a sequence, failing to check a required parameter, or not responding to a notification. Omission errors reflect gaps in the human-machine communication that should have prompted the missing action — feedback signals that were absent, insufficiently salient, or too easily overlooked.

Systemic Causes of Communication Errors

Individual communication errors are typically symptomatic of systemic conditions that create vulnerability to that type of error:

Interface complexity that exceeds users' cognitive capacity creates chronic conditions for all types of error — users who are cognitively overloaded make more input errors, miss more feedback signals, and are more likely to misinterpret complex outputs.

Feedback inadequacy that leaves users without accurate information about system state creates systematic conditions for mode errors, omission errors, and undetected errors — users acting on incorrect or incomplete situational awareness.

Design-user model mismatch creates systematic conditions for mode errors and interpretation errors — users whose mental models of the system diverge from the system's actual design will regularly act on incorrect beliefs about system state and will regularly misinterpret system outputs.

High-stakes time pressure creates conditions for all error types — users under time pressure take cognitive shortcuts, attend less carefully to feedback signals, and are more likely to execute practiced patterns automatically rather than deliberating about the specific requirements of the current situation.

Error Analysis and Prevention

Understanding human-machine communication errors requires analyzing both the immediate cause — the specific failure in input, processing, or output that constituted the error — and the contributing conditions that made the error possible or likely. Effective error prevention targets contributing conditions rather than just immediate causes: modifying interface designs that systematically produce certain error types, rather than attempting to train users out of cognitive patterns that interact badly with the current design. A key principle is that human cognitive limitations are relatively fixed, and designs that require users to exceed those limitations will produce predictable errors regardless of training and intention.