Human errors are now the leading cause of cybersecurity incidents in organizations. Contrary to popular belief, these vulnerabilities are not always the result of malicious intent, but often stem from psychological or environmental factors such as stress, fatigue, or impaired concentration. A recent study proposes an innovative solution to anticipate these risks in real time.<\/p>\n
The idea is based on a system capable of continuously analyzing users’ physiological and cognitive states, as well as their immediate environment. Sensors measure indicators such as heart rate, skin electrodermal response, ambient temperature, noise levels, and lighting. This data, often imperceptible to the users themselves, reveals signs of stress or mental overload before errors are even made. For example, a temperature increase beyond 24 degrees or noise exceeding 55 decibels significantly reduces concentration and decision-making abilities. Similarly, overly intense or poorly adapted lighting can cause visual and mental fatigue, thereby increasing the risk of oversight or mishandling.<\/p>\n
The core of the system uses an artificial intelligence model combining two technologies: a convolutional neural network to identify patterns in sensor data, and a recurrent neural network of the LSTM type to track the evolution of these patterns over time. This approach enables the detection of subtle changes in the user’s state, such as rising stress or a drop in attention, with 84% accuracy. When a high risk is identified, the system can trigger appropriate actions, such as a warning, temporary access restriction, or an alert to a supervisor.<\/p>\n
One of the major challenges lies in the ability to generalize this model across different individuals. Each person reacts differently to stress or their environment. To address this, the system was designed to adapt to each user without requiring complete retraining. It uses dynamic data normalization, allowing it to distinguish between normal variations and early signs of a potential error. However, this method may have limitations if the initial data does not cover a sufficiently wide range of situations.<\/p>\n
Another issue concerns privacy protection. Physiological data, such as heart rate or skin response, can reveal sensitive health information, such as pregnancy, mental disorders, or chronic illnesses. Their collection and use therefore raise ethical and legal questions. In many countries, such data is considered sensitive and subject to strict regulations. Organizations wishing to deploy such a system must therefore ensure they obtain informed consent from employees, guarantee transparency regarding data usage, and implement clear policies to prevent any discrimination or breach of privacy.<\/p>\n
The system was tested using simulated data replicating real-world conditions. The results show that it can identify at-risk states with good reliability, although improvements are still possible, particularly to reduce false alerts. An individual calibration phase, where each user would wear the device for a full day, could further refine accuracy by accounting for each person’s specific characteristics.<\/p>\n
In practice, this type of technology could transform how organizations approach cybersecurity. Instead of focusing solely on technical solutions like firewalls or antivirus software, it places humans at the center of the protection strategy. By adapting security measures based on users’ actual states, it offers a proactive approach to preventing errors before they occur. This could not only reduce the number of incidents but also improve employee well-being by avoiding extreme stress or cognitive overload situations.<\/p>\n
Human errors are now the leading cause of cybersecurity incidents in organizations. Contrary to popular belief, these vulnerabilities are not always the result of malicious intent, but often stem from psychological or environmental factors such as stress, fatigue, or impaired concentration. A recent study proposes an innovative solution to anticipate these risks in real time.<\/p>\n
The idea is based on a system capable of continuously analyzing users’ physiological and cognitive states, as well as their immediate environment. Sensors measure indicators such as heart rate, skin electrodermal response, ambient temperature, noise levels, and lighting. This data, often imperceptible to the users themselves, reveals signs of stress or mental overload before errors are even made. For example, a temperature increase beyond 24 degrees or noise exceeding 55 decibels significantly reduces concentration and decision-making abilities. Similarly, overly intense or poorly adapted lighting can cause visual and mental fatigue, thereby increasing the risk of oversight or mishandling.<\/p>\n
The core of the system uses an artificial intelligence model combining two technologies: a convolutional neural network to identify patterns in sensor data, and a recurrent neural network of the LSTM type to track the evolution of these patterns over time. This approach enables the detection of subtle changes in the user’s state, such as rising stress or a drop in attention, with 84% accuracy. When a high risk is identified, the system can trigger appropriate actions, such as a warning, temporary access restriction, or an alert to a supervisor.<\/p>\n
One of the major challenges lies in the ability to generalize this model across different individuals. Each person reacts differently to stress or their environment. To address this, the system was designed to adapt to each user without requiring complete retraining. It uses dynamic data normalization, allowing it to distinguish between normal variations and early signs of a potential error. However, this method may have limitations if the initial data does not cover a sufficiently wide range of situations.<\/p>\n
Another issue concerns privacy protection. Physiological data, such as heart rate or skin response, can reveal sensitive health information, such as pregnancy, mental disorders, or chronic illnesses. Their collection and use therefore raise ethical and legal questions. In many countries, such data is considered sensitive and subject to strict regulations. Organizations wishing to deploy such a system must therefore ensure they obtain informed consent from employees, guarantee transparency regarding data usage, and implement clear policies to prevent any discrimination or breach of privacy.<\/p>\n
The system was tested using simulated data replicating real-world conditions. The results show that it can identify at-risk states with good reliability, although improvements are still possible, particularly to reduce false alerts. An individual calibration phase, where each user would wear the device for a full day, could further refine accuracy by accounting for each person’s specific characteristics.<\/p>\n
In practice, this type of technology could transform how organizations approach cybersecurity. Instead of focusing solely on technical solutions like firewalls or antivirus software, it places humans at the center of the protection strategy. By adapting security measures based on users’ actual states, it offers a proactive approach to preventing errors before they occur. This could not only reduce the number of incidents but also improve employee well-being by avoiding extreme stress or cognitive overload situations.<\/p>\n
DOI:<\/strong> https:\/\/doi.org\/10.1186\/s42400-026-00609-z<\/a><\/p>\n Title:<\/strong> BioEnvSense: a human-centred security framework for preventing behaviour-driven cyber incidents<\/p>\n Journal:<\/strong> Cybersecurity<\/p>\n Publisher:<\/strong> Springer Science and Business Media LLC<\/p>\n Authors:<\/strong> Duy Anh Ta; Farnaz Farid; Farhad Ahamed; Ala Al-Areqi; Robert Beutel; Tamara Watson; Alana Maurushat<\/p>\n","protected":false},"excerpt":{"rendered":" Can Artificial Intelligence Prevent Human Errors in Cybersecurity? Human errors are now the leading cause of cybersecurity incidents in organizations. Contrary to popular belief, these vulnerabilities are not always the result of malicious intent, but often stem from psychological or environmental factors such as stress, fatigue, or impaired concentration. A recent study proposes an innovative… Continue reading Can Artificial Intelligence Prevent Human Errors in Cybersecurity?<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10,7],"tags":[],"class_list":["post-30","post","type-post","status-publish","format-standard","hentry","category-environment","category-health","entry"],"_links":{"self":[{"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/posts\/30","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/comments?post=30"}],"version-history":[{"count":1,"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/posts\/30\/revisions"}],"predecessor-version":[{"id":31,"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/posts\/30\/revisions\/31"}],"wp:attachment":[{"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/media?parent=30"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/categories?post=30"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/socialsciencesreview.com\/en\/wp-json\/wp\/v2\/tags?post=30"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}