The rise of Artificial Intelligence (AI) has fundamentally changed the security world. Whereas AI was previously seen primarily as a tool to automate complex tasks, it increasingly plays a dual role.
Security vendors have recently fully committed to integrating AI into their solutions. This gives them the opportunity to better detect and prevent attacks. A model can be trained to recognize suspicious traffic in order to block a potential attack and initiate further investigation. The accuracy of this pattern recognition is rapidly increasing due to the amount of available data on which the models are trained.
As beneficial as the progression of security solutions is, on the other hand, lurks the danger of cybercriminals professionalizing attacks using AI. Instead, they seek ways to make attack detection more difficult. Thereby, the mass adoption of AI applications also introduces whole new risks. Orange Cyberdefense research shows that security professionals face the business and technical implications of AI now and in the future.
Orange Cyberdefense’s Security Navigator 2025 offers companies insights into the latest trends so they can adjust their strategy. Orange Cyberdefense bases these insights on data from its Security Operations Centre, supplemented by other sources. Based on this, the report maps the threat landscape as accurately as possible and outlines an outlook for the year ahead. This allows organizations to better prepare for the modern threat landscape.
Battle of two camps to new heights
The good news is that organizations are increasingly relying on AI to bolster their security strategy. Modern intrusion detection systems (IDS) and endpoint security solutions are powered by models that spot anomalous behaviour in networks and systems. These security tools thus analyze and prevent threats in real time. The models can also help identify phishing attempts and block malware based on behavioral characteristics rather than known signatures. Detection without signatures is also needed to discover new forms of malware, which are being developed at lightning speed these days.
Artificial Intelligence can quickly process large amounts of data and recognize patterns in them. For human security analysts, those suspicious patterns often remain invisible simply because viewing and assessing everything is impossible. In a modern security strategy, there is certainly a place for AI, but the technology is far from perfect. For example, the current state of AI raises questions about the reliability and security of the systems, especially when they unintentionally make wrong decisions.
The biggest risk, however, comes from amplified cyberattacks. One of the most common applications is using AI to generate compelling phishing emails and deepfake videos. These types of attacks play on human emotions and vulnerabilities. Not all targets know how to deal with this well, which increases the likelihood of being taken advantage of.
In addition, there are developments such as “dark large language models (LLMs).” These models are designed for malicious purposes. They can generate code for malware, identify system vulnerabilities, and even automate social engineering campaigns. The accessibility of AI technology means that even less tech-savvy criminals can carry out sophisticated attacks.
Pressure on organizations
The widespread mass adoption of AI by companies also brings business challenges. Many organizations feel fear of being left behind and the pressure to embrace AI solutions. As a result, AI implementation has exploded in recent years, often without sufficient attention to the potential risks. The risk of not implementing solutions or implementing them too late weighs heavily, but the reality is that rapid adoption can lead to data breaches and unforeseen vulnerabilities. Moreover, many AI systems are still in their infancy. LLMs’ training relies on datasets that may not be fully representative or secure. As a result, models may contain unintended biases or vulnerabilities that attackers exploit.
AI’s dual role in cybersecurity and broad business impact show that a proactive and balanced approach is essential. Companies are best advised to invest in AI-driven security technology while remaining vigilant to the threat of AI-driven attacks. In this way, AI can hopefully work best for the defense side of cybersecurity.
Are you curious about more insights from Security Navigator 2025? Then check out the full report.
Tip: Explosive increase in cyber extortion puts entire supply chain at risk