Are Security Teams Keeping Pace With AI-Accelerated Threats?

A recent BBC article reports on a call from more than 100 technology and cybersecurity companies, including Google, Microsoft and OpenAI, to strengthen global cyber defenses.

The underlying open letter argues that AI-enabled attacks will become more widespread and sophisticated as models grow more capable. It also points to the weaknesses likely to leave organizations exposed: longstanding bugs, excessive permissions, misconfigurations, weak authentication and unpatched systems.

These are established cybersecurity problems but AI changes how quickly and widely they can be targeted.

The real question for security leaders is what that shift means for the people responsible for finding weaknesses, testing defenses, and responding to attacks.

AI can help an attacker move faster through parts of an existing workflow. It may support reconnaissance, explain unfamiliar code, adapt a script or help analyze the results of an attempted exploit.

The impact will vary significantly according to the target, vulnerability, tools available, and expertise of the person using them. AI assistance does not automatically turn a novice into an experienced operator. It can, however, reduce the time someone with existing knowledge spends on individual tasks.

This distinction matters. Completing a task faster is different from operating effectively across an unfamiliar environment.

An attacker still needs to decide which lead is worth pursuing, understand how systems connect and adapt when an approach fails. On the defensive side, security professionals still need to determine which findings matter, assess the potential impact and choose a response appropriate to the environment.

AI can accelerate parts of that work but technical judgment holds the work together.

Security professionals already work with tools that automate discovery, analysis and response. AI extends that pattern by making more information and potential actions available more quickly. More output does not always produce better decisions.

• A vulnerability researcher needs to evaluate whether an apparent weakness is genuinely exploitable. 
• A detection engineer must understand whether suspicious activity reflects malicious behavior or legitimate use. 
• Application security, cloud security, threat hunting and incident response teams all need to interpret findings in the context of the systems they protect.

Those decisions rely on practical experience. People who understand operating systems, networks, applications, identities, and attack paths are better equipped to use AI effectively because they can challenge its output. They can recognize when a recommendation lacks context, introduces risk, or simply does not work.

AI therefore increases the value of cybersecurity expertise. As more parts of the workflow accelerate, organizations depend even more on people who can direct the work and make sound decisions when the answer is unclear.

AI can be a valuable learning and operational tool when it is used with a human in the loop. It can explain concepts, suggest approaches, and help practitioners work through unfamiliar material.

AI can become an increased risk when cybersecurity professionals use it as a shortcut. If someone follows an AI-generated command without understanding what it does, they may struggle to validate the result or adjust when the environment responds differently than expected.

Another risk, overreliance, can create gaps in independent problem-solving when foundational skills do not develop alongside AI-assisted workflows.

Security training should account for both. Practitioners need experience using new tools, while still being challenged to investigate problems, test assumptions, and recover from failed approaches themselves. That is how they develop the judgment to know when AI is helping and when it is leading them in the wrong direction.

When attacks move faster, organizations need people who can apply their knowledge without waiting for a familiar playbook.

That capability is developed through practice. Security professionals need opportunities to work in realistic environments where the answer is not immediately obvious. They need to follow an attack path, interpret what they find, and adjust their approach as conditions change.

Practical assessments help professionals demonstrate their ability to apply what they have learned. They also help leaders identify where teams need further development before those gaps are exposed during an incident.

OffSec helps organizations strengthen the technical understanding, practical experience and specialist expertise their teams need as AI accelerates cybersecurity work.

• Build practical foundational capability. Learn Enterprise gives organizations access to hands-on education across penetration testing, security operations, incident response, cloud security, secure development and other disciplines. Learners develop skills by applying concepts in labs, while organizations can create learning paths that reflect the capabilities different roles require.
• Test skills in realistic environments. OffSec Enterprise Cyber Range scenarios give teams space to practice offensive and defensive work against active threats. Practical assessments and certifications provide additional evidence that learners can apply their knowledge, adapt their approach and work through unfamiliar problems.
• Develop specialized AI security expertise. AI also introduces systems and attack surfaces that require focused knowledge. AI-300 and the OSAI certification prepare professionals to test LLMs, AI agents and supporting infrastructure. The AI Red Teaming Upskill Program provides a structured path for organizations developing this capability across a team.

The distinction is important. Every security team needs the practical foundations to operate effectively in AI-accelerated workflows. Organizations building or deploying AI systems may also need specialists who understand how those systems can be attacked.

The strongest response to AI-driven change is a workforce that can use faster tools without surrendering the judgment behind the work.

Security leaders should give their teams room to build that judgment now, through hands-on practice that challenges them to investigate, adapt and make decisions for themselves. AI can help them move faster. Their expertise determines whether they move in the right direction.