Digital overload can weaken vigilance and facilitate manipulations, according to an analysis devoted to technostress and cognitive sovereignty, signed by Eric Farro, expert at Orange Group, and reproduced by ZATAZ. Technostress, i.e., stress linked to intensive use of digital tools, thus joins the human factors that social engineering, which aims to lead a person to adopt a specific behavior, exploits to provoke a risky decision.
- Digital overload weakens vigilance and makes manipulation easier.
- Social engineers exploit urgency, fatigue and the difficulty of verifying requests.
- Automation shifts humans toward supervision, but it does not reliably reduce their cognitive load.
- Tools that estimate mental state raise privacy risks, so local processing and limited centralization are needed.
Overload prepares the ground for manipulation
A notification interrupts a search, another forces a switch to a different app, a message demands an immediate reply and a technical alert awaits validation. Each of these solicitations may seem justified. Their accumulation, however, fragments reasoning and complicates priority hierarchy.
Eric Farro describes an imbalance between human cognitive capacities and information environments that constantly demand of their users. The working memory, which temporarily retains the information needed for a task, holds a limited amount of items. Attention selects those that will be processed. When demands exceed these resources, understanding of the situation can degrade.
This limit directly affects the field of social engineering. The cybercriminal seeks to obtain information, the opening of a document, the acceptance of a connection or the execution of an operation. To achieve this, they can rely on fatigue, urgency and the difficulty of verifying several elements at once. The approach channel remains an email, a call, a messenger or a physical interaction, while overload creates the conditions in which the pretext becomes more effective.
The analysis takes the example of an illustrative scenario: a person receives several requests presented as priority, then an interlocutor posing as IT support offers to “fix the problem”. The promise of restoring normal operation may encourage a hasty validation. This case describes a possible mechanism and does not correspond to an incident reported in the analysis.
Prevention therefore also concerns work organization. A verification procedure must remain feasible during a busy period, because if each check adds a disproportionate difficulty, the temptation to shortcut the process increases. Interfaces, validation circuits and interruption management thus contribute to resistance against manipulations.
Automation shifts pressure and risks
Autonomous agents, systems capable of chaining certain operations on their own, can take over part of the work. The human then becomes a supervisor of decisions or actions whose steps they no longer necessarily follow all. Eric Farro notes that this evolution transforms cognitive load without guaranteeing it decreases.
The supervisor must understand the results, spot anomalies and know when to intervene. Typically reliable automation can nevertheless reduce active vigilance. When an unusual event occurs, one must quickly reconstruct the context and decide under pressure, a situation in which a malicious interlocutor could invoke a tool failure, an urgent exception or a validation presented as indispensable.
Technologies also attempt to estimate fatigue or mental load. The analysis mentions in particular pupillometry (measurement of pupil size variations), heart‑rate variability, electrodermal activity (variations in skin conductance) and certain keyboard‑typing characteristics. The envisaged goal is to adapt assistance, by simplifying information, modifying control sharing or putting a system into safety mode. A question immediately arises: who can access these vulnerability estimates, and for what purposes?
Eric Farro also describes a theoretical risk for supervision centers. An actor controlling both the alert flow and the estimation of analysts’ cognitive state could seek to cause saturation at the opportune moment. Protection therefore concerns the collected data, the inferences produced and the actions that may follow. Local processing, limiting centralization and separating uses are among the avenues examined.
