News

signageOS Adds AI Security Scans for 180000 Screens

Digital Signage Trends Editorial
Digital Signage Trends Editorial
· 4 min read
Developer at a desk in a software office at dusk reviewing code for digital signage security

On 18 November 2026, device management company signageOS said it has joined Anthropic’s Cyber Verification Program. It now uses the Claude Mythos AI model to look for vulnerabilities in the cloud services, APIs and software behind more than 180000 connected screens on over 50 hardware platforms. For screen owners, this raises the bar for what you should expect from any signage vendor.

Key takeaways

  • signageOS now runs Anthropic’s Claude Mythos to find security bugs in software that manages more than 180000 signage devices.
  • The AI scans add to signageOS’s existing SOC 2 Type II and ISO 27001 certifications, annual penetration tests and required multi-factor login.
  • Vendor-side AI scanning does not protect against weak passwords, shared logins or screens on guest Wi-Fi. Those are still your job.
  • Use this news as a reason to ask your CMS vendor five security questions this quarter.

What did signageOS announce?

signageOS is using Anthropic’s restricted Claude Mythos model to check its own code for security holes. invidis reported that the scans cover its cloud infrastructure, APIs and device management software, and that checks run on software releases before they ship.

The AI sits on top of controls the company already had: SOC 2 Type II and ISO 27001 certifications, a yearly independent penetration test (a paid outside attempt to break in) and required multi-factor authentication. CTO Michael Žabka put the reasoning simply: “The platform that manages a screen network has to be at least as secure as the network itself.”

What is Claude Mythos and who can use it?

Mythos is Anthropic’s most capable model for security work, and only vetted organizations can use it. Anthropic released Claude Mythos 5.2 on 9 November 2026 and gives access through verification programs. On 6 October it rebuilt its Cyber Verification Program to widen access for security teams, and enrolled companies must accept data retention so Anthropic can monitor for misuse.

In practice, a signage software company can now point a model at a large codebase and get findings that used to take a human auditor weeks.

Why does this matter if you run screens?

The real target is the management layer, not the TV. A CMS (content management system, the cloud software that pushes content to your screens) controls every screen at once, so one weak point can expose all of them.

Hand plugging a cable into a network switch to separate signage screens from guest Wi-Fi

This autumn showed how that plays out. We covered the German QSR menu board hack, where attackers changed displayed prices, and the XCPE botnet on Android players.

Two cases show the difference:

  • A 12-site burger chain with 36 menu boards. One compromised admin login means wrong prices in every store during the lunch rush. Vendor code quality and login controls matter most here.
  • A clinic with 3 waiting room screens. The risk is local: players sitting on the same Wi-Fi patients use. Network setup matters more than vendor AI.

Will vendor AI scanning make your screens safe?

No. It reduces bugs in the vendor’s code, which is good, but it cannot see your network or your staff habits. The German QSR case started with a default password, a problem no code scanner fixes. Expect AI audits to become a standard vendor talking point in 2027, and judge vendors on the basics too.

What should you ask your signage vendor this quarter?

Ask five questions and write down the answers. Our guide to choosing a CMS provider covers the wider checklist.

QuestionGood answerRed flag
Do you hold SOC 2 Type II or ISO 27001?Yes, report shared under NDA"In progress" with no date
When was your last outside penetration test?Within 12 monthsNo date given
Can admins enforce MFA for all users?Yes, by policyOptional per user
How are player updates checked before release?Automated scans, staged rolloutManual pushes to all screens
How fast do you patch critical bugs?Written target in daysNo stated policy

What should you do on your own network now?

Fix the basics that attackers actually use. This takes an afternoon for most networks under 20 screens.

  1. Turn on multi-factor login for every CMS user and remove former staff.
  2. Change default passwords on players and commercial displays.
  3. Move screens off guest Wi-Fi onto a separate VLAN (a walled-off section of your network).
  4. Keep player operating systems patched. Our note on Windows players after Patch Tuesday shows how updates can reboot screens mid-day, so schedule them overnight.
  5. Name one person who owns the screens. Unowned networks are the ones nobody audits.

FAQ

What is signageOS?

signageOS is a digital signage device management company. Its software supports more than 180000 connected devices across over 50 hardware platforms, according to invidis.

Is my screen network safer because of this?

Only if your CMS runs on signageOS, and only on the vendor side. Your passwords, user accounts and network setup still decide most of your risk.

Can any company use Claude Mythos?

No. Anthropic limits Mythos to organizations it verifies through its Cyber Verification Program or Life Sciences program. Enrolled companies must accept data retention for misuse monitoring.

Should a small business with 2 screens worry?

Yes, but the fix is cheap. Use a unique password with MFA, keep screens off guest Wi-Fi and remove logins for people who have left.

What certifications should a signage vendor have?

SOC 2 Type II and ISO 27001 are the two most common. Ask for the date of the last independent penetration test as well.

Sources