AI is not creating a fresh cybersecurity crisis so much as compressing the timeline on an old one: the hidden flaws in legacy software are now being found faster than many teams can fix them.
Software flaws do not disappear when teams move on; they can linger, travel through suppliers, and resurface as security problems in SCADA, AI-assisted coding, and other exposed systems.
Dragos’s move into Project Glasswing points to a new kind of security testing: frontier models probing OT software before weak code can become an operational problem.
Verizon’s latest breach trend data points to software vulnerabilities overtaking stolen credentials, while AI is helping attackers move from discovery to exploitation in hours.
The initiative is framed as a defensive workflow for finding software vulnerabilities and helping counter cyber threats, but its real significance is how tightly AI must be controlled before it can be trusted inside security operations.
A fresh security update wave across Adobe’s creative and commerce products puts patch timing, version inventory, and exposure management back at the center of defensive strategy.
Daybreak appears to be more than a standalone model name: it signals an AI cyber offering built around GPT-5.5, with defensive workflows and tighter access controls for sensitive security tasks.
As Anthropic’s Mythos exposes decades-old software flaws, boards and CISOs scramble to respond with AI-powered defenses.
Recent fixes in the popular Spring framework reveal the ongoing, high-stakes battle against hidden software vulnerabilities.
U.S. cybersecurity officials sound alarms as new waves of attacks exploit overlooked vulnerabilities in widely used business software.
A new breed of AI models is exposing thousands of software flaws at lightning speed-leaving users, companies, and even global banks scrambling to keep up.
Federal agencies and companies face a race against time as hackers target critical Fortinet, Microsoft, and Adobe vulnerabilities now confirmed as actively exploited.
New research reveals advanced AI models routinely invent or overlook critical software vulnerabilities, quietly embedding risk into modern development.
Critical vulnerabilities lurking in Mozilla software have been patched, closing doors to cybercriminals and reinforcing user trust.
For the first time, an autonomous AI agent weaponized human language to breach major GitHub repositories, leaving tech giants scrambling.
Archipelo and Checkmarx join forces to inject developer identity and workflow context into the heart of application security.
A staggering majority of organizations are running exposed software as outdated code, risky libraries, and poor security practices fuel a new wave of cyber threats.
Hackers are weaponizing software flaws at unprecedented speed, leaving defenders scrambling to keep up.