AI in the enterprise: a hypervitaminized brother-in-law who demands control

Artificial intelligence (AI) has burst forcefully into companies, but its adoption without control can become a double-edged sword. Sergio Martínez, SonicWall's regional director for Iberia and Italy, sums it up with a metaphor that leaves no one indifferent: "AI is a hypervitaminized brother-in-law". With this image, Martínez warns that AI, like that relative who has an opinion on everything without having a clue, "what it doesn't know, it makes up". And in the business context, that can translate into information leaks, wrong decisions or security breaches.

In an interview with Computerworld, Martínez reviews the evolution of the CISO's role in a scenario dominated by AI. His diagnosis is clear: AI is a powerful tool both for defending and for attacking, but cybercriminals have the advantage. "AI is a very good tool for defending, but above all for attacking," he acknowledges. And he adds: "There is nothing we can do about that. Only prepare ourselves, because they are going to attack us from anywhere".

The CISO, a key player in business strategy

Given this panorama, the CISO's role takes on crucial relevance. Martínez maintains that the CISO must be part of the management committee, not only as head of security, but as a strategic actor who helps assess the risks that affect the company's survival. "The risk is no longer solely IT-related, but affects the survival of the company itself," he stresses. The nature of attacks has changed: previously technical vulnerabilities were exploited, now identity theft and credentials have become the new perimeter. "Credentials are the new perimeter, which is where the battle is," he states.

For SMEs, this reflection is especially relevant. Many do not have a dedicated CISO, but they should designate someone responsible for security and privacy, with a voice in strategic decisions. AI can help identify weak points, but also help attackers find them. Hence the importance of investigating alerts: "More than half of critical alerts are never investigated," Martínez warns. A worrying figure that reveals the lack of resources and attention that many organizations devote to security.

Obsolete VPNs and Zero Trust: an urgent transition

Another of the critical points Martínez addresses is traditional VPNs. According to his estimates, more than half of current attacks occur through obsolete VPNs, especially the SSL VPNs deployed during the pandemic and that have never been patched. "They are a source of problems," he states. Many manufacturers have already withdrawn them from their catalogs, and SonicWall encourages replacing them with zero-trust access models (ZTNA).

The Zero Trust philosophy is based on maximum distrust and minimum access: "Tell me who you are and I'll tell you what you can access". But it is not enough to verify the user's identity; you also have to evaluate the device from which they connect, its security status and the connection conditions. This approach is more secure than traditional VPNs, which allow access to all internal network resources once authenticated. "That is outrageous," Martínez admits.

For an SME, migrating to ZTNA may seem like a challenge, but there are cloud solutions that facilitate the transition and allow controlling access to on-premises, IaaS and SaaS environments, in addition to filtering Internet traffic and regulating the use of AI tools. In fact, Martínez mentions that their proposal allows blocking unauthorized AI tools and prioritizing corporate solutions such as Copilot or Claude.

Recommendations for SMEs and IT teams

In light of Martínez's warnings, at ForgeNEX we recommend that SMEs and IT teams take immediate measures:

• Establish AI usage policies: it is not about prohibiting, but about regulating. Define which tools are allowed, what data can be interacted with and train employees on the risks.
• Review and update the remote access infrastructure: replace obsolete SSL VPNs with ZTNA solutions that apply the principle of least privilege.
• Investigate security alerts: allocate resources to analyze critical events and do not let them pass. Automation with AI can help, but it requires human oversight.
• Incorporate security into business strategy: even without a full-time CISO, designate a person responsible who participates in key decisions and assesses risks.
• Protect credentials: implement multi-factor authentication and properly manage privileged access.

AI is not a passing fad, but neither is it a magic wand. As Martínez recalls, it is a "hypervitaminized brother-in-law" who can help us or get us into a mess. The key is to use it with judgment and, above all, with security. In an environment where almost 90% of SMEs suffered ransomware attacks last year, there is no room for improvisation.

To delve deeper into how AI is transforming cybersecurity, you can consult our analysis on AI and geopolitics in cybersecurity investment or the article on GPT-6 Cyber.

Source: Computerworld. Analysis and adaptation: ForgeNEX.