AI is advancing rapidly but IT security is lagging behind

Unlimited opportunities, but also immense risks

Artificial intelligence (AI) is radically transforming the way organizations operate, promising unprecedented efficiency, automation, and scalability. However, there is a clear disconnect between the speed at which AI is advancing rapidly, IT security is lagging behind, and the ability of IT infrastructures to keep up with cybersecurity. According to a new HelpNetSecurity report published on July 30, 2025, this gap is becoming a major vulnerability.

What the 2025 AI and Security Training Report Highlights

The report was compiled based on responses from hundreds of technology and IT security professionals around the world. The data shows that despite growing trust in AI, only 36% of organizations say they have a clear security strategy for AIFurthermore, the processes and policies governing AI within companies are generally incomplete or incipient.

The main findings of the report

• 64% of companies admit that they do not have sufficient knowledge to manage the risks generated by AI.
• Only 29% have comprehensive AI governance policies in place.
• 51% of IT leaders say their organization is using AI to improve security, but the lack of a robust compliance framework raises critical questions.

These numbers point to a simple but dangerous truth: technology is evolving faster than our ability to secure it.

Real risks associated with the lack of an AI security strategy

Although implementing AI may seem beneficial in the short term, without a well-defined security strategy, companies expose themselves to more sophisticated attacks, even based on AI technologyVulnerable AI systems can be compromised to:

• give wrong or manipulated answers;
• filter sensitive data to attackers through “prompt injection”;
• detect and exploit weaknesses in networks automatically;
• propagates malware invisibly through automated channels.

Adversarial AI model – an emerging danger

Adversarial AI refers to techniques by which artificial intelligence models are manipulated to produce errors. For example, Seemingly harmless images or text can be designed to influence the outcome of an AI model, which can lead to wrong decisions, intrusions or unauthorized access.

Why are organizations lagging behind in securing AI?

There are several barriers to adopting a solid security strategy in the AI era:

1. Lack of specialized expertise

AI and cybersecurity are complex fields, each with its own technical requirements. A major task facing organizations is recruiting or training experts who can understand both areas.

2. Lack of regulatory frameworks or clear standards

In the absence of clear and up-to-date industry standards, Organizations lack clear direction to assess or mitigate risks associated with AI.

3. Rushing to adopt AI before assessing security implications

Often, Organizations prioritize the business benefits of AI, sacrificing the audit steps, risk analysis, and implementation of security controls.

What should companies do to close this gap?

The solutions are not simple, but the next steps can be decisive for the future of any organization that integrates AI into operational flows.

1. Implementing clear and transparent AI governance

Every implemented AI system must have clear policies regarding:

– how input data is collected, stored and analyzed;

– who is responsible for auditing algorithms and AI ethics;

• – performance evaluation criteria with an emphasis on integrity and security.

2. Periodic AI risk assessment

A security audit dedicated to AI systems must be performed with the same rigor as applied to classic IT systems. Penetration testing, line-of-code analysis, and verification of the integrity of machine learning strategies are essential.

3. Adapting the security strategy according to the evolution of AI

IT security must be agile, able to adapt to the emergence of new technologies, including generative AI or LLMs (Large Language Models). Automating the detection and response process (XDR, modern SIEM) becomes crucial for anticipating AI-driven attacks.

AI in cybersecurity – enemy or ally?

Paradoxically, AI can be both a threat and a solution in the field of IT security.

Strategic uses of AI in security

• Detection of network anomalies in real time;
• User Behavior Analysis (UBA);
• Attack prediction based on historical patterns;
• Automating incident response.

However, for AI to help and not to make vulnerable, a cyber architecture is needed that understands the operating mechanisms of AI technologies and be able to proactively intervene in case of deviation or compromise.

Conclusion: It's time for organizational maturity in AI

Looking to the future, it is clear that organizations that manage to align AI innovation with rigorous security standards will have a clear competitive advantageBut for this, they must:

• invest in people and specific AI + security skills;
• build a model of ethical and transparent governance;
• adopt an organizational culture focused on constant assessment of technological risks.

The gap between AI evolution and IT security maturity is not only a technical issue, but also a strategic one. Anyone who fails to keep up exposes the entire organization to a digital storm that is difficult to predict.

You have certainly understood what the news in 2025 is related to cybersecurity, if you are interested in deepening your knowledge in the field, we invite you to explore our range of courses dedicated to cybersecurity in the category CybersecurityWhether you're just starting out or want to brush up on your skills, we have a course for you.