Data study ·
AI speeds up cyberattacks and puts traditional training under pressure
Original title, in French: “L’IA accélère les cyberattaques et met les formations traditionnelles sous pression”
- Publication
- Alliancy (France)
- Author
- Tiago Gil
- Published on
- 28 September 2026 · read in the page structured data
Summary
According to SoSafe's report "Les leaders de la cybersécurité en action", drawn from a summit in Paris attended by 150 cybersecurity managers, 76% of French cyber managers see attackers actively using AI tools, notably for vishing and scraping of professional data. Internally, more than 90% of professionals now use AI at work, compared with 10% two years earlier, giving rise to "shadow AI" that escapes IT control. The report points to the limits of traditional awareness training: 37% of IT managers have no reliable feedback on its effectiveness. SoSafe recommends measuring real behaviours, as the company STIME does, receiving more than 1,000 reports of suspicious emails per month.
Summary based on the full text of the article.
Read the article on Alliancy →Figures cited
-
76 %
French cyber managers seeing active use of AI by attackers
France, cybersecurity managers
« 76 % des responsables cyber français constatent une utilisation active d’outils d’IA par les attaquants. »
Source: SoSafe, rapport « Les leaders de la cybersécurité en action »
-
80 %
share of attacks that are vishing
Observed attacks
« Le vishing, ou hameçonnage vocal, représenterait désormais 80 % des attaques observées. »
Source: SoSafe
-
90 %
professionals using AI at work
Professionals, now vs two years ago
« plus de 90 % des professionnels reconnaissent désormais utiliser l’IA au travail, contre 10 % deux ans auparavant »
Source: SoSafe
-
37 %
IT managers with no reliable feedback on training effectiveness
IT managers
« 37 % des responsables informatiques déclarent par ailleurs ne disposer d’aucun retour fiable sur l’efficacité réelle de leurs programmes de formation »
Source: SoSafe
-
85 %
CIOs finding regulatory guidelines too complex
CIOs
« 85 % des DSI jugent les directives réglementaires trop complexes à diffuser en interne. »
Source: SoSafe
-
1,000 emails
emails reported each month at STIME
STIME
« plus de 1 000 mails sont ainsi signalés chaque mois après la mise en place d’un dispositif reposant sur un geste de signalement simple et un retour systématique aux collaborateurs »
Classification
- Country
- France
- Company size
- All sizes
- Industries
- Software vendors, IT services companies, telecoms · All industries, or industry not specified
- Companies
- SoSafe (study author), STIME (deploys AI)
- Keywords
- cybersecurity, shadow AI, phishing, employee training, vishing
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 Alliancy and its author. Entry translated from French by Claude Sonnet 5.5.