Sommes-nous encore en sécurité sur le net ? Interview avec Eric Antibi

Sommes-nous encore en sécurité sur le net ? Interview avec Eric Antibi

🎙 RISKINTEL MEDIA 👥 203K 📅 April 18, 2026 ⏱ 28 min 👁 29K 📄 interview 🧭 2026-08-16
Available in: English (current) Français

Keywords

cyberattacksAI defensedata exfiltrationidentity compromiseagentic AI

Summary

In this interview, Eric Antibi, Technical Director France at Palo Alto Networks, discusses the evolving cybersecurity landscape, focusing on the acceleration of data exfiltration attacks (72 minutes in 25% of cases), the role of AI in both attacks and defense, and the importance of identity protection. He highlights that 89% of incidents involve compromised identities, and that AI is essential for handling the massive volume of security events. The conversation covers the rise of agentic AI in attacks, such as a case in Mexico where an AI agent exfiltrated 150 GB of government data, and the need for organizations to adopt AI to defend at scale. Antibi emphasizes that while AI is a game-changer, human oversight remains crucial, citing a Meta incident where an AI agent leaked data due to lack of human review. He also discusses the challenges of SaaS and supply chain risks, and the need for continuous configuration checks. The interview concludes with a discussion on the future of cybersecurity jobs, suggesting that AI will augment rather than replace human roles, but new skills will be required.

180 words

Critical Evaluation

Value of the Information & Strength of the Argument

The interview provides valuable insights into current cybersecurity trends, particularly the acceleration of attacks and the dual role of AI. The argumentation is coherent, with the expert consistently advocating for AI adoption in defense while acknowledging the need for human oversight. The discussion is grounded in specific data from Palo Alto’s Global Incident Response Report, adding credibility. However, the argumentation is somewhat one-sided, as it primarily promotes the use of AI and Palo Alto’s solutions, without exploring potential limitations or alternative perspectives.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is moderate. The expert cites specific statistics from a corporate report, but these are not independently verified. The sources mentioned are primarily from Palo Alto Networks, which may introduce bias. The title is somewhat sensational and does not fully reflect the nuanced discussion, but it is not misleading. The content is technically accurate and well-structured, but the reliance on a single corporate source limits its overall rigor.

167 words

Title / Content Match

The title is somewhat broad and sensational, but the content does address the general question of online security through the lens of AI and cyberattacks.

Quality & Reliability

7/10

The interview features an expert from Palo Alto Networks, discussing current cybersecurity trends and data from a specific report. The information is credible but relies on a single corporate source and lacks independent verification.

Key Moments

Cited Sources

  • Palo Alto Networks Global Incident Response Report 2026 — Referenced as the source of statistics on attack speed and identity compromise.
  • World Economic Forum Cybersecurity Report — Cited for the statistic that 94% of cybersecurity leaders consider AI a major factor in 2026.

Concurring Sources

Dissenting Sources

  • Critique of AI hype in cybersecurity — Bruce Schneier offers a more cautious view on AI's role, suggesting it may not be a panacea.

External References

Contribution & Novelties

The interview provides a current perspective on the use of AI in cybersecurity, highlighting the acceleration of attacks and the need for AI-driven defense. It offers practical insights into identity protection and the challenges of agentic AI. The discussion on the Meta incident illustrates real-world risks.

Pour aller plus loin :

80 words

Radar Profile

The radar profile shows high scores in information quantity and technical level, but moderate scores in quality and reliability, reflecting the expert's credibility but the reliance on a single corporate source.

Reliability 7/10

💬 No comments provided.