Cyberattacks boosted by AI are exploding: faster, autonomous, and accessible, they give the advantage to attackers. To keep up, companies must also adopt defenses that are agents themselves.
The landscape of AI-related threats is evolving at a frantic pace. With Anthropic’s recent discovery of the first AI-driven cyber espionage campaign, we now have concrete evidence of what many security actors feared. Last year, I hypothesized that LLMs offered attacked companies an asymmetric advantage, but this balance of power has changed. The emergence of agentive AI and the development of offensive infrastructures have paved the way for malicious actors to operationalize large-scale chains of agentive tools. And at the current time, they have the advantage.
The impact is already visible in France. According to the French Ministry of the Interior, 348,000 cyber incidents were recorded in 2024, representing a 74% increase compared to the previous five years. Everywhere in France, attacks are multiplying and becoming more targeted and difficult to detect.
This situation is not surprising. Recent initiatives have shown that qualified offensive operators training agents specifically for threat hunting can outperform individual researchers. Similar capabilities are now available to malicious actors. They now have a roadmap to use AI to execute multi-stage attacks autonomously, without being limited by human intervention.
More vulnerabilities are being exploited
AI agents have significantly reduced the time between the discovery of a vulnerability and its exploitation.
Recently, Google announced that its Big Sleep project had identified numerous zero-day vulnerabilities in open-source projects. A collaboration between DeepMind and Project Zero, Big Sleep included a set of multi-phase agents designed to identify software vulnerabilities and develop functional exploits.
Although Big Sleep allowed security managers to prevent these exploits from materializing, there is no doubt that malicious actors are using the same techniques to compromise their targets. For French organizations subject to NIS2 directive requirements, the speed at which vulnerabilities now turn into exploits increases the stakes in terms of compliance deadlines and incident reporting obligations.
Attackers chain agents
It is no longer just a theory: attackers break down attack phases into distinct agentic workloads and use chains of agents to execute each phase autonomously.
Anthropic’s report on cyber espionage revealed that Chinese malicious actors used AI agents to carry out 80 to 90% of attacks independently. Human intervention was required less than seven times at critical decision points. AI agents, executing thousands of requests per second, significantly reduced the time and human resources needed for an attack.
Additionally, in its 2025 Threat Intelligence Report, Anthropic reveals that AI allows less skilled malicious actors to learn and execute more advanced tactics, techniques, and procedures. Cybercriminals with minimal technical expertise, for example, used Claude to develop and sell several ransomware variants for 345 to 1,029 euros on Internet forums. They relied solely on AI to implement encryption algorithms and evasion techniques.
Thanks to AI, it has never been cheaper for attackers to arm exploits. Agentive AI offers greater autonomy to these attacks, which are on the rise. This trend is expected to lead to a proliferation of attacks targeting companies’ most valuable data.
To counter this threat, attacked companies must respond with equally agentive defense strategies.
Fighting agents with agents
As attackers develop chains of agentic tools, internal red teams and defense teams must also increase their own use of agentive AI. They need AI agents that leverage internal system resources to gain context, then break down defensive tasks into workloads to speed up vulnerability identification and remediation.
For effective execution, these agents must have a deep understanding of the software environment. An infrastructure that provides the right data and context to deployed agents is necessary, such as knowledge graphs that map relationships across source code.
When given access to knowledge graphs, agents can combine company knowledge with historical data on vulnerabilities and known security anti-patterns to help teams prioritize threats based on real attack models rather than theoretical risks.
In addition to prevention, agentive defenses also promote resilience. Companies can break tasks down into detection and correction activities within their organization’s runbooks. Agents handle everything from identification to investigation, correction, and post-mortem analysis to reduce downtimes and limit damage.
These use cases illustrate the measures companies can take to strengthen their agentive defenses. The convergence of technical capabilities and processes offers new tools to combat malicious actors and develop defensive operations. Attackers are taking advantage of these tools, so it is now time for companies to follow suit.
