Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

My two previous recent postings on AI covered “Agentic AI” and how that impacts cybersecurity and the eventual emergence of malicious agentic AI malware. Both of those articles started to touch on the idea of automated agentic AI defenses. This posting goes into a little more detail on what agentic AI defenses might mean. It starts with agentic AI, which is a collection of automated programs (i.e., bots or agents) working toward a common goal.

Our recent research reveals a concerning discrepancy between employees' confidence in their ability to identify social engineering attempts and their actual vulnerability to these attacks. While 86% of respondents believe they can confidently identify phishing emails, nearly half have fallen for scams in the past. This disconnect between perceived competence and demonstrated vulnerability, the "confidence gap", poses a substantial risk to organizations. The Danger of Overconfidence.

A phishing campaign is impersonating travel agency Booking.com to target employees in the hospitality industry, according to researchers at Microsoft. The attacks use a social engineering technique called “ClickFix” to trick victims into downloading malware.

The rise of connected devices has fundamentally reshaped industries, enabling unprecedented levels of automation, efficiency, and innovation. These devices fall under the Extended Internet of Things (XIoT), a broad category encompassing traditional Internet of Things (IoT) devices, operational technology (OT), industrial control systems (ICS), the Internet of Medical Things (IoMT), and other connected assets that span enterprise IT and operational environments.

Today, one of the greatest challenges that cyber defenders face is analyzing detection hits from indicator feeds, which provide metadata about specific indicators of compromise (IOCs), like IP addresses, ASNs, domains, URLs, and hashes. While indicator feeds have proliferated across the threat intelligence industry, most feeds contain no contextual information about why an indicator was placed on the feed.

Security and attacks continues to be a very active environment, and the visibility that Cloudflare Radar provides on this dynamic landscape has evolved and expanded over time. To that end, during 2023’s Security Week, we launched our URL Scanner, which enables users to safely scan any URL to determine if it is safe to view or interact with.

In 2024, we announced Log Explorer, giving customers the ability to store and query their HTTP and security event logs natively within the Cloudflare network. Today, we are excited to announce that Log Explorer now supports logs from our Zero Trust product suite. In addition, customers can create custom dashboards to monitor suspicious or unusual activity.

With cybercriminals becoming increasingly more sophisticated, utilizing the latest tools such as generative AI and SaaS exploits, the cybersecurity world in 2025 appears to be more convoluted than ever before. From compromised credentials-driven SaaS attacks to social engineering-based fraud facilitated through deepfakes, CISOs need to remain ahead with insightful recommendations.

With the rapid and dynamic nature of the digital world of today, businesses are seeing a mounting high rate of cybersecurity attacks. Cyber attackers keep evolving and coming up with new methods of breaching their systems, which leaves security teams under immense pressure to identify, assess, and remediate vulnerabilities at scale. Traditional methods of vulnerability management are typically behind the curve because the sheer volume of threats is overwhelming.

On March 10, 2025, X (formerly Twitter) experienced a series of outages due to a large-scale Distributed Denial of Service (DDoS) attack. The platform went down multiple times throughout the day, affecting millions of users globally. Elon Musk later confirmed that X was targeted by a "massive cyberattack", with initial investigations pointing to Ukraine-based IP addresses as the source.