Is AI Overhyped in Cybersecurity? Insights from Picus Labs Red Report 2025

Artificial Intelligence (AI) is often seen as the next big frontier in cybersecurity, with many speculating that it could revolutionize how cybercriminals operate. However, a recent analysis by Picus Labs in their Red Report 2025 paints a more tempered picture. After examining over a million malware samples, the report highlights that AI is not yet the game-changer many expect it to be, at least not in the way it’s commonly portrayed in the media. Despite this, cybercriminals continue to evolve, using well-established tactics and techniques to carry out their attacks.

Key Findings

The Red Report 2025 reveals some notable cybersecurity trends. Contrary to the buzz surrounding AI-driven cyberattacks, there’s been no significant surge in AI-based tactics in 2024. While adversaries are leveraging AI to improve the effectiveness of traditional methods—like creating more convincing phishing emails or improving malicious code—they have not yet fully harnessed AI’s potential in their attacks. Cybersecurity teams are advised to focus on known attack techniques, as they remain dominant.

One of the most alarming findings is the sharp rise in credential theft, with attacks targeting stored passwords and cached logins increasing more than threefold. This surge emphasizes the need for robust credential management and continuous threat detection. The report also highlights that a majority of malware attacks rely on a core set of MITRE ATT&CK techniques, such as process injection and application layer protocols, which can evade detection if not properly monitored.

Despite the sophistication of modern attacks, the Red Report underscores that defense teams should focus on the fundamentals. Strengthening defenses against credential theft, improving threat detection, and conducting continuous security validation can be more effective than chasing the latest AI-driven threat hype.

What Undercode Say:

In today’s cybersecurity landscape, it’s easy to get caught up in the hype surrounding emerging technologies, especially artificial intelligence. The media portrays AI as the next frontier in both offensive and defensive cyber tactics. However, the reality, as highlighted by the Picus Labs’ Red Report 2025, is that AI is still not playing a major role in cyberattacks. While it’s true that adversaries are beginning to use AI for specific tasks—like crafting more convincing phishing emails or automating malicious code generation—the core tactics used in most attacks today remain largely unchanged.

The increase in credential theft is one of the most concerning trends highlighted in the report. Cybercriminals are targeting password stores and cached logins with alarming frequency, using them to escalate privileges and move laterally through networks. This uptick shows that organizations must prioritize secure credential management and employ proactive threat detection mechanisms. The stealthy nature of modern malware, especially infostealers, makes these attacks even harder to detect, as they often operate under the radar, blending in with regular network traffic.

Interestingly, while AI might have the potential to change the way cybercriminals operate in the future, the attacks we’re seeing today rely on tried-and-true techniques that have been in use for years. The MITRE ATT&CK framework, which tracks common attack methods, shows that adversaries are primarily using a small set of techniques to achieve their objectives. These include process injection, which allows attackers to hide malicious activities within legitimate system processes, and using application layer protocols like HTTPS or DNS-over-HTTPS for command-and-control and data exfiltration.

These techniques are difficult to detect through traditional signature-based methods because they blend seamlessly with normal network traffic. This is why security teams should focus on behavioral analysis rather than relying solely on signature detection. By monitoring how attacks unfold and correlating suspicious activities, defenders can spot potential threats more effectively, even if they’re masked by seemingly legitimate processes.

One key takeaway from the Red Report is that cybersecurity teams should not be distracted by the AI hype. While AI will undoubtedly play a larger role in the future, the most pressing threats today are based on a narrow set of attack techniques. Organizations need to focus on strengthening their defenses around these known threats, especially credential protection, threat detection, and continuous security validation.

In an era where cybersecurity budgets and resources are often limited, prioritizing these fundamental security practices can make a world of difference. Organizations should implement robust credential management policies, monitor for suspicious behavior across their networks, and invest in solutions that continuously validate the effectiveness of their security controls. By sticking to these essentials, security teams can significantly reduce their exposure to the most common and damaging types of attacks.

Moreover, while the media loves to focus on AI as the ultimate weapon of cybercriminals, the reality is that cybersecurity has always been a battle between defenders and attackers using evolving techniques. The rise of AI might change the game eventually, but for now, the focus should remain on defending against attacks that are happening today, using the tools and techniques that have been proven effective.

The Red Report 2025 also stresses the importance of a mindset shift. Rather than chasing the next big thing, cybersecurity professionals should focus on continuous improvement, refining their skills, and ensuring their security systems are up to date. By doing so, they can better prepare for the challenges that lie ahead, whether they involve AI or more traditional attack methods.

In conclusion, while

References:

Reported By: https://thehackernews.com/2025/02/debunking-ai-hype-inside-real-hacker.html
https://www.instagram.com
Wikipedia: https://www.wikipedia.org
Undercode AI: https://ai.undercodetesting.com