Keeping WordPress websites secure in the AI era
WordPress security in the AI era means patching faster, trusting fewer plugins and checking what every connected tool is allowed to do, because attackers now find and use flaws in a fraction of the time they used to. A website that gets compromised can lose its search visibility along with its data, which is why security work sits alongside the SEO for B2B companies on WordPress that Priority Pixels delivers. We cover technical SEO as part of that service, alongside search experience optimisation and conversion rate optimisation.
The latest sign of the shift came from the WordPress project itself. As Search Engine Journal reported in its coverage of the new WordPress security initiative, the core team has launched a programme to scale up security work as AI makes vulnerabilities easier to find and exploit. A few days earlier, a popular SEO plugin was found creating login credentials for its own support tool without telling website owners.
Both stories point the same way for anyone responsible for a WordPress website. What has changed, where the risk now sits and the practical steps that keep a WordPress website secure are set out below.
What the WordPress core security initiative is
The WordPress Core Security Initiative is a programme from the WordPress security team to handle a sharp rise in security reports, much of it caused by better AI models. The announcement on the WordPress security blog describes a “substantial increase in the volume of incoming security reports” and says much of this growth reflects “the rapid advancement of frontier AI models”.
The team has set out three areas of work. Each one tackles a different part of the problem, from how fixes are shipped to how flaws are found in the first place.
A better release process
A tighter and more automated security release process with improved testing. Security releases will also be scheduled in advance.
Breaking the backlog
Working through open security findings with the aim of bringing them down to zero. The volume of reports had grown faster than the team could clear them.
Finding flaws with AI
Scanning and tooling assisted by AI to find vulnerabilities before they can be exploited. The same technology attackers use is turned on the code first.
The initiative covers WordPress core, the software every WordPress website runs on. Plugins and themes sit outside it. That is where most of the reported problems come from.
Scheduled security releases are a useful change for anyone who manages updates. Knowing when a release is due means testing and deployment time can be planned rather than found at short notice.
Why AI is changing WordPress security
AI is changing WordPress security because it speeds up both halves of the work, finding a flaw and turning it into an attack. The National Cyber Security Centre’s assessment of the impact of AI on the cyber threat to 2027 says AI will almost certainly make parts of cyber intrusion more effective and efficient. It also warns that “the time between disclosure and exploitation has shrunk to days and AI will almost certainly reduce this further”.
The NCSC expects a digital divide between systems that keep pace with these threats and a large proportion that are more vulnerable. For WordPress websites the dividing line comes down to how quickly updates are applied and how many plugins are installed.
Research from the WordPress security company Patchstack shows the scale of the problem. Its report on the state of WordPress security in 2026 counted 11,334 new vulnerabilities in the WordPress ecosystem in 2025, a 42% increase on 2024. Of those, 46% did not receive a fix from the developer in time for public disclosure and the weighted median time to first exploit was 5 hours.
The same report found that the large majority of new vulnerabilities sit in plugins, with a small share in themes and very few in WordPress core. It also recorded a significant rise in poor vulnerability reports generated by AI, which adds to the workload of every team that has to triage them.
The WordPress security team has responded to that noise directly. Its update to the WordPress vulnerability disclosure programme says that for assets other than core and the block editor, issues that need a role granted by an administrator are generally no longer eligible unless they show a high severity escalation. Researchers are encouraged to focus on flaws that can be exploited without logging in or by Subscribers.
What the WordPress 7.0.2 release shows about patching speed
The WordPress 7.0.2 release shows that serious core flaws still happen and that the project will now push fixes out automatically when the risk is high enough. The WordPress 7.0.2 security release fixed one critical and one high severity issue, including a REST API flaw that could lead to remote code execution. Because of the severity, the WordPress.org team enabled forced updates for websites running affected versions and released fixed versions 6.9.5 and 6.8.6 for older branches.
A website that has switched off automatic updates or is stuck on an unsupported version does not benefit from that safety net. On the timescales Patchstack describes, even a few days of delay is long enough to matter.
The weeks that followed brought several more developments, set out in order below. Each one changes how WordPress websites should be looked after.
July 2026
WordPress 7.0.2 security release with forced updates
28 August 2026
Core Security Initiative announced
30 August 2026
Rank Math support agent access reported
1 September 2026
Disclosure programme rules tightened
Taken together, these events show the core project raising its game while the wider plugin ecosystem remains the weaker point. That makes the choice and management of plugins the biggest lever most website owners have.
Keeping the server software current matters too. Our guide to keeping PHP up to date explains why the version of PHP a website runs on carries its own security risk.
Plugins are where most of the risk sits
Plugins are where most WordPress risk sits because each one adds code written by a different developer, with its own update record and its own access to the website. A website running dozens of plugins depends on dozens of separate teams patching quickly, which many of them fail to do.
The WordPress.org detailed plugin guidelines set some ground rules for plugins in the official directory. Guideline 7 bars plugins from contacting external servers without explicit consent, which matters more now that plugins increasingly connect to AI services.
Every plugin removed is one less thing to patch and one less route in. Our guide to WordPress plugin bloat covers how to review an installation and cut it back without losing features the website needs.
- Remove plugins you no longer use.
- Choose plugins with a recent update history.
- Test updates and apply them within days.
- Check what access each plugin asks for.
- Keep inactive plugins installed.
- Install plugins their developers have abandoned.
- Leave updates waiting for months.
- Switch on remote support features without reading what they do.
A shorter list of plugins, each kept current, is easier to secure than a long list that nobody reviews. It also loads faster, which helps both visitors and search performance.
What the rank math support agent episode means for website owners
The Rank Math episode shows that a trusted plugin can grant itself access to a website without the owner noticing. According to Search Engine Journal’s report on the Rank Math allegations, version 1.0.277 of the plugin created a WordPress Application Password when a connected administrator opened its Help and Support area. The plugin runs on over 4 million websites and the password inherited the role of whoever opened that screen.
Rank Math has since paused the feature. Its statement on pausing the support agent says version 1.0.277.2 removes it, that the agent could not go beyond the user’s role and that it should have told users more clearly. The company says the feature will return with an explicit request for access.
Open Users, then Profile, for every administrator and check the Application Passwords section. Revoke any password you do not recognise or no longer use, including any labelled as the Rank Math Support Agent.
Whatever the intent, the episode is a useful prompt to check which credentials exist on a website. Application Passwords were added in WordPress 5.6 and the Application Passwords integration guide explains that they are listed on each user’s profile, where they can be revoked. A Last Used column shows which ones are still active.
The same check is worth repeating every few months. Integrations, reporting tools and former suppliers can all leave credentials behind long after they stop being needed.
Plugin settings screens deserve the same scrutiny. Any feature that offers remote support or an AI assistant should explain what access it needs before it is switched on.
New risks when AI tools connect to WordPress
AI tools connected to WordPress bring new risks because they can be manipulated through the content they read as well as through the code they run. The NCSC guidance on AI and cyber security describes prompt injection attacks as “one of the most widely reported weaknesses in LLMs”. An attacker crafts input that makes the model behave in a way nobody intended, such as revealing confidential information or triggering actions in a connected system.
On a WordPress website, that might be a chatbot plugin with access to customer data or an AI writing assistant that can publish content. If either can be steered by text it reads, an attacker does not need to break into the website to misuse it.
The practical answer is to give AI tools the least access they need. A content assistant rarely needs an administrator account and a chatbot should never be able to read data it has no reason to show a visitor.
How security problems affect search visibility
Security problems affect search visibility because Google warns searchers away from websites it believes have been hacked. Google’s help page on the Security issues report in Search Console says affected pages can appear with a warning label in search results or an interstitial warning page in the browser.
Hacked websites are also used to host spam pages, inject links or redirect visitors elsewhere, all of which damage rankings long after the clean up. Regular technical SEO checks catch many of these signs early, such as unexpected pages in the index or sudden changes in crawl patterns.
UK obligations for website security
Organisations in the UK have a legal duty to protect personal data held on their websites. The ICO’s guide to data security explains that UK GDPR requires appropriate technical and organisational measures and points to basic technical controls such as those in Cyber Essentials. Contact forms, account areas and WooCommerce stores all hold personal data that falls under these rules.
Cyber Essentials is the scheme backed by the Government that sets out the minimum standard of cyber security for organisations, built around five technical controls. Priority Pixels is Cyber Essentials certified, the UK Government security certification scheme backed by the National Cyber Security Centre.
Many public sector and healthcare buyers now ask suppliers for Cyber Essentials before work begins. Checking that a website partner holds it is a quick way to judge how seriously they treat security.
Practical steps to keep a WordPress website secure
Keeping a WordPress website secure in the AI era comes down to faster updates, fewer plugins, tighter access and constant monitoring. The WordPress documentation on plugin and theme automatic updates says that to keep a website secure you should always update plugins and themes to the latest version.
The table below sets out the main areas to cover and what good practice looks like in each. Most of these can be checked in an afternoon.
| Area | What to check | Why it matters now |
|---|---|---|
| Core updates | Automatic security updates switched on and the website on a supported version | Forced security updates only help websites that can receive them |
| Plugins and themes | Unused plugins removed and the rest updated after testing | Most new vulnerabilities are found in plugins |
| User accounts | Few administrators, strong passwords and two factor authentication | Access that is never granted cannot be misused |
| Application Passwords | Every password recognised and the unused ones revoked | Plugins and integrations can create credentials without telling you |
| AI tools | Each tool limited to the access it needs | Prompt injection can turn a helpful tool against the website |
| Monitoring and backups | Malware scanning, vulnerability alerts and tested backups | Problems are found in hours rather than weeks |
Updates need testing before they reach the live website, because a broken form or checkout costs leads too. The balance is to test quickly and deploy within days, not to hold updates back for months.
Hosting plays a part as well. A managed host that applies security hardening and keeps daily backups gives a website a better starting point. The managed WordPress hosting page sets out what that involves.
Access reviews belong in the same routine. Removing accounts for staff and suppliers who have moved on closes one of the easiest routes into any website.
Ongoing WordPress maintenance and monitoring
Ongoing WordPress maintenance keeps security work from slipping, because updates, scans and access reviews happen on a schedule rather than after a problem. Priority Pixels combines continuous malware scanning with monitoring of core and plugin vulnerabilities through its own reporting platform, so threats get actioned, not just logged.
Every update is tested in a staging environment first, with compatibility checks and a rollback plan. Urgent incidents, such as a website being offline or compromised, receive a response within two hours during working hours and are worked until resolved. The full service is set out on the WordPress maintenance and security services page.
With AI shortening the gap between a flaw being found and being used, regular care is what keeps a WordPress website on the right side of the divide the NCSC describes. Security and search visibility depend on each other, so both belong in the same plan.
FAQs
How secure is WordPress?
WordPress core is maintained by a dedicated security team and serious flaws are fixed quickly, sometimes through forced automatic updates. Most problems come from plugins, themes, weak passwords and websites that are not kept up to date.
How do I protect my WordPress website from hackers?
Keep core, plugins and themes updated, remove plugins you do not use and limit the number of administrator accounts. Add two factor authentication, review Application Passwords regularly and run malware scanning with tested backups.
How can I tell if my WordPress website has been hacked?
Common signs include unfamiliar administrator accounts, pages you did not create appearing in search results, unexpected redirects and warnings in Google Search Console. A security scan and a review of recently changed files will usually confirm it.
Does the WordPress Core Security Initiative cover plugins?
No, the initiative focuses on WordPress core, the software every WordPress website runs on. Plugins and themes remain the responsibility of their developers, so website owners still need to vet and update them.