A recent ZDNET report puts a striking price tag on a once-expensive skill: roughly $4 for an AI model to help find a zero-day flaw. The report describes rapidly advancing AI models making it cheaper and easier to uncover critical vulnerabilities, and calls the trend a double-edged sword and a gift to hackers. For everyday users, the practical question is what AI zero-day vulnerability discovery changes about staying safe. The short answer: the time you can wait to update is getting shorter.
What AI-Driven Zero-Day Discovery Actually Means
A zero-day is a software flaw that the vendor does not know about yet, so no fix exists when attackers first find it. Historically, finding one took specialized skill, time, and money. That scarcity kept many of them in the hands of well-resourced researchers, criminal groups, and state-backed teams.
According to ZDNET, advancing AI models are lowering that barrier. The headline claim is that vulnerability discovery can now cost a few dollars, which means even mediocre hackers can potentially profit from it. The source summary is brief, so it is worth being careful about what it does and does not establish: it points to cheaper, easier discovery, not to a guarantee that every AI-found flaw becomes a working attack.
The same capability helps defenders too. Security teams can use these models to find and fix bugs before criminals do. That is the double edge: the technology is neutral, and the outcome depends on who uses it first and how quickly fixes reach users.
Why Cheaper Exploits Shrink Your Patching Window
When discovery was expensive, there was often a comfortable gap between a flaw being found and it being widely abused. Cheap discovery pressures that gap from both sides. More people can look for bugs, and they can look more often.
For you, the key moment is the time between a vendor releasing a patch and your device installing it. During that period, the flaw is publicly known but your system is still exposed. If attackers can find and test flaws faster and at lower cost, that period becomes more dangerous, so leaving updates pending for weeks is a riskier habit than it used to be.
This does not mean panic is warranted. Most people are not individually targeted with zero-day attacks. But cheap tooling tends to widen the pool of attackers, and broader pools tend to favor opportunistic campaigns against whoever has not updated.
Where AI Tools Themselves Become the Attack Surface
There is a second angle that often gets overlooked. AI is not only a tool for finding bugs in other software; AI products are software too, and they can have flaws of their own. Assistants, browser agents, and extensions often have broad access to your accounts, files, and web sessions, which makes them attractive targets.
A concrete example is research from Zenity, covered in our report on zero-click flaws in Claude and ChatGPT Atlas. Zenity disclosed two zero-click AI browser hacking techniques capable of hijacking Anthropic's Claude in Chrome extension and OpenAI's ChatGPT Atlas browser. The lesson is that the more authority you hand to an AI tool, the more a single flaw in it can matter.
If cheap AI-assisted discovery is also pointed at AI software, the tools you installed for convenience deserve the same scrutiny as any other app with deep access.
What This Means For You
You do not need to become a security expert, but a few habits matter more as the patching window narrows:
- Speed matters more than perfection. Installing updates promptly is the single most effective response to faster flaw discovery.
- Fewer tools means less exposure. Every extension, plugin, and AI assistant is another piece of software that may contain an undiscovered flaw.
- Permissions are risk. An AI tool that can read your email, browse as you, or access files can do real damage if hijacked.
How to Reduce Your Exposure as a Regular User
- Turn on automatic updates for your operating system, browser, phone, and apps. This closes the gap between a patch being released and you receiving it.
- Restart when prompted. Many updates do not take effect until you reboot or relaunch the browser.
- Audit your browser extensions. Remove any you no longer use or do not recognize, and be cautious with those requesting broad access to all sites.
- Review your AI tools. Check what accounts, files, and browsing sessions each assistant can reach, and limit permissions to what you actually need.
- Use strong, unique passwords and multi-factor authentication so one compromised app does not unlock everything else.
- Retire unsupported devices. Hardware and software that no longer get patches cannot respond to newly found flaws.
The Bottom Line
The ZDNET report highlights a shift that is easy to state and hard to ignore: AI zero-day vulnerability discovery is getting cheap, and that benefits attackers as well as defenders. You cannot control how fast flaws are found, but you can control how fast you patch and how much access your tools have. Start today by enabling automatic updates and reviewing the AI tools and browser extensions you rely on. For a real-world example of how AI software can itself be exploited, read our coverage of the Zenity zero-click findings.




