In September 2026, at WordCamp US, security researcher Aaron D. Campbell described a recent WordPress Core vulnerability report he had to print out and mark up with a pen just to follow the logic. It was eleven pages long. Eleven pages of chained steps that no human would have discovered or exploited on their own. An AI agent found it, validated it, and wrote scripts to automate the exploit — all for about twenty-five cents in compute.

That story stuck with me because it captures something I've been watching happen in slow motion across the WordPress ecosystem. The security game hasn't just gotten harder. It's changed categories entirely.

The old threat model is dead

For years, WordPress security advice followed a predictable pattern. Keep your plugins updated. Use strong passwords. Install a security plugin. Don't use admin as your username. That advice still matters, but it's no longer sufficient because the threat model has fundamentally shifted.

The old model assumed attackers were humans with limited time and attention. A vulnerability might exist in a plugin for months before someone stumbled onto it. Exploit development took days or weeks. There was a gap between discovery and attack, and that gap was your patching window.

That gap is gone.

Campbell, who ran the WordPress Security Team and now works at Monarx (a malware detection company that white-labels security for hosting providers), described the new reality plainly: AI agents can now chain together vulnerabilities that would require an eleven-page document to explain. They operate twenty-four hours a day. The moment a vulnerability is found, exploitation can begin automatically. There is no breathing room.

What AI-powered attacks actually look like

This isn't theoretical. The pattern is showing up in real incidents across the WordPress ecosystem.

An AI agent doesn't just find a single vulnerability in a plugin and exploit it. It finds a minor information disclosure in one plugin, uses that to learn about the site's configuration, chains that with an authentication bypass in a different plugin, and then escalates privileges through a third vector that a human security researcher would never have connected. Each individual step might look innocuous. The chain is devastating.

The volume is also unprecedented. Campbell's company, Monarx, operates across multiple hosting providers, which gives them a cross-platform view of attack patterns. They're seeing new malware variants appear in one host's ecosystem and spread to others before traditional detection can catch up. The speed of propagation has compressed from days to hours.

And the motivation, as always, is money. Ransomware, crypto mining, credential theft, phishing infrastructure — the same criminal playbook, now executed at machine speed with machine creativity.

What actually works now

Campbell was careful not to leave the audience in despair. The situation is overwhelming, he said, but not hopeless. The path forward involves several shifts in how we think about WordPress security.

Automatic updates are non-negotiable. The patching window has shrunk from weeks to hours. If you're still manually reviewing plugin updates before applying them, you're leaving a window open that AI-powered attackers can walk through. WordPress core's automatic update mechanism exists for a reason. Use it for plugins too.

Choose hosts that invest in security tooling. Not all hosting is equal, and most cheap hosts do the bare minimum. Ask your host what they run behind the scenes. Companies like Monarx operate as white-label security layers that many hosts bundle into their plans without advertising it. A host that takes security seriously will have answers when you ask.

Monitor for compromised credentials. If your admin password has appeared in any breach database, AI agents will find it and try it. Credential stuffing at machine speed is trivially easy. Use unique passwords and enable two-factor authentication on every WordPress admin account.

Reduce your plugin surface area. Every active plugin is a potential attack vector. Deactivate and delete anything you're not actively using. The math is simple: fewer plugins mean fewer places for AI agents to find a foothold.

Watch for supply chain attacks. The WordPress plugin ecosystem has seen a rise in attacks that compromise a legitimate plugin update and push malicious code to thousands of sites at once. This is harder to defend against, but keeping your plugin count minimal and sticking to well-maintained plugins with active development teams helps.

The arms race isn't new, but the speed is

Campbell drew an interesting parallel to the history of cryptographic hashing. When the computing power to crack older hash algorithms became cheap enough, the industry had to adapt. The same thing is happening with vulnerability discovery and exploitation. AI is the new computing power advantage, and the security community is in the adaptation phase.

The encouraging part is that defenders have access to AI too. Security teams are using the same class of tools to find vulnerabilities before attackers do, to detect anomalous behavior faster, and to automate responses. The arms race continues, but the weapons on both sides have gotten significantly more powerful.

For the average WordPress site owner, the practical takeaway is straightforward: the basics still matter, but the urgency has increased dramatically. What used to be "update your plugins this week" is now "update your plugins today, because something is already scanning for the vulnerability you're sitting on."

The eleven-page vulnerability report is the new normal. Your job is to make sure your site isn't the one it applies to.

Source: WP Tavern Jukebox Podcast #232 — Aaron D. Campbell on Navigating WordPress Security in the AI Era, recorded at WordCamp US, September 2026.