Data study ·
In French companies, 88% of cybersecurity leaders report unauthorised use of AI
Original title, in French: “Dans les entreprises françaises, 88 % des responsables cyber constatent des usages non autorisés de l’IA - Le Journal des Entreprises”
- Publication
- Le Journal des Entreprises (France)
- Author
- Unsigned
- Published on
- 2 September 2026 · read in the page structured data
Summary
According to Okta's Global CISO Insights 2026 report, produced with Apprize360 Intelligence and based on 306 cybersecurity leaders in six countries (France, United Kingdom, Germany, United States, Canada, Japan), 88% of French respondents say they have observed unauthorised use of AI by their employees. The study also highlights the difficulties posed by autonomous AI agents: only 47% of respondents can identify all the agents present in their IT environment, and 81% are concerned about the excessive access rights granted to AI. France shows the lowest confidence in the oversight of AI agents among the countries studied, while 36% of UK organisations report having advanced, automated governance. AI governance is becoming a matter for senior management, with only 31% of leaders saying they are aligned with their leadership on the acceptable level of risk.
Summary based on the full text of the article.
Read the article on Le Journal des Entreprises →Figures cited
-
88 %
share of French cybersecurity leaders who have observed unauthorised use of AI
France, sample of approximately 34 cybersecurity leaders, 2026
« En France, 88 % des responsables de la cybersécurité interrogés par Okta indiquent avoir constaté au moins une utilisation non autorisée de l'IA par leurs collaborateurs. »
Source: Global CISO Insights 2026, Okta
-
47 %
share of leaders able to identify all AI agents in their environment
six countries studied, 306 respondents
« seuls 47 % des responsables interrogés déclarent pouvoir identifier l'ensemble des agents IA présents dans leur environnement informatique »
Source: Global CISO Insights 2026, Okta
-
46 %
share of leaders able to control the resources accessible to AI agents
six countries studied
« Ils sont 46 % à pouvoir contrôler les ressources auxquelles ces agents peuvent accéder »
Source: Global CISO Insights 2026, Okta
-
81 %
share of leaders concerned about excessive access rights granted to AI
six countries studied
« 81 % se disent préoccupés par les droits d'accès excessifs accordés à l'intelligence artificielle »
Source: Global CISO Insights 2026, Okta
-
36 %
share of UK organisations reporting advanced, automated governance of AI agents
United Kingdom
« 36 % des organisations interrogées déclarent disposer d'une gouvernance avancée et entièrement automatisée de leurs agents IA »
Source: Global CISO Insights 2026, Okta
-
31 %
share of leaders fully aligned with their senior management on acceptable AI risk
international, six countries
« seulement 31 % des responsables cyber interrogés déclarent être totalement alignés avec leur direction générale et leur conseil d'administration sur le niveau de risque lié à l'IA qu'ils jugent acceptable »
Source: Global CISO Insights 2026, Okta
Classification
- Country
- France, United Kingdom, Germany, United States, Canada, Japan
- Company size
- All sizes
- Industries
- Software vendors, IT services companies, telecoms · All industries, or industry not specified
- Companies
- Okta (study author), Apprize360 Intelligence (cited)
- Keywords
- cybersecurity, shadow AI, AI agents, AI governance, identity management, okta
Entry written by Marsanne Studio from the cited article, with the help of an AI model (claude-sonnet-5). The rights to the article and its illustrations belong to Le Journal des Entreprises. Entry translated from French by Claude Sonnet 5.5.