Your smart TV, your video doorbell, and the thermostat you installed three winters ago all have something in common: they quietly sit on your home network, and most of us never think about them again. According to a Financial Times report published today, October 13, 2026, that "set it and forget it" habit could become a much bigger problem in the age of autonomous AI.
Anthropic, the company behind the Claude AI models, has warned that future AI agents could behave like self-replicating worms, hunting for poorly secured connected devices and spreading from one to the next. Importantly, this is a warning about what could happen, not a report of an outbreak already underway. Still, it is a useful wake-up call, and the fixes are surprisingly simple.
What Exactly Is an "AI Worm"?
A computer worm is malware that copies itself from one machine to another without anyone clicking a thing. Classic worms follow a fixed script: scan for a known weakness, exploit it, repeat. That makes them noisy and, eventually, predictable. Security teams learn the pattern and write defenses.
An AI-driven version would be different. Instead of following a rigid script, it could adapt as it goes, trying new approaches when the first one fails, adjusting to different device types, and working around obstacles. Think of the difference between a vending machine and a determined problem-solver. That flexibility is what worries researchers.
Why Your Home Gadgets Are an Easy Target
Attackers do not need to break into your laptop if a forgotten device offers an easier way in. Connected home products tend to share a few weaknesses:
- Neglected updates. Many devices never receive a software update after the day they are unboxed.
- Default passwords. Cameras and routers that still use factory credentials are low-hanging fruit.
- Short support windows. Budget gadgets can stop getting security patches within a couple of years, while still working fine for a decade.
- Flat home networks. On most home Wi-Fi, every device can talk to every other device, so one weak link can expose the rest.
The number of connected devices per household keeps climbing, too. Apple, for example, is widely expected to unveil new smart-home hardware at its "Welcome Home" event today, and Amazon and Google already sell a steady stream of displays, speakers, and cameras. More convenience means more doors to lock.
9 Steps to Lock Down Your Smart Home
You do not need to be a security expert. Work through this checklist over a weekend and you will be ahead of most households.
- Turn on automatic updates. For every device that offers the option, enable it. Then manually check the ones that do not.
- Replace default passwords. Use a unique, long password for each device and for your router's admin page. A password manager makes this painless.
- Create a separate network for smart devices. Most modern routers support a guest or IoT network. Put TVs, cameras, and appliances there, and keep laptops and phones on the main one.
- Use modern Wi-Fi encryption. Choose WPA3 if your router supports it, or WPA2 at minimum.
- Disable features you do not use. Remote access, UPnP, and voice-assistant integrations you never touch are all unnecessary exposure.
- Check your router's firmware. The router is the front door to your whole network, so keep it current, or replace it if the manufacturer has stopped issuing updates.
- Retire unsupported gear. If a device no longer receives security patches, either disconnect it from the internet or replace it.
- Buy with support in mind. Before purchasing, look for a stated update commitment. A slightly pricier device with five years of patches beats a bargain gadget with none.
- Review what is connected. Open your router's device list. If you spot something you do not recognize, investigate it.
The Bigger Picture: AI Cuts Both Ways
It would be easy to read this story as purely alarming, but the same technology that could power smarter attacks can also strengthen defenses. Security teams already use AI to spot unusual behavior, triage alerts, and find vulnerabilities before criminals do. Today's warning is part of a broader conversation about building safeguards early, rather than scrambling after the fact.
For everyday users, the takeaway is refreshingly practical: the basics matter more than ever. Strong passwords, timely updates, and a segmented network will not stop every threat, but they make you a far harder target, and attackers, human or automated, tend to move on to easier ones.
Stay Informed
Cybersecurity news moves quickly, and the line between hype and genuine risk is not always obvious. For more plain-English coverage, browse our cybersecurity section, see what readers are following on our trending page, or explore the latest in AI news and gadgets.
Frequently Asked Questions
Are AI worms spreading through homes right now?
No. Anthropic's warning, as reported by the Financial Times, concerns a possible future threat. There is no confirmed outbreak of autonomous AI worms in consumer devices today.
Which devices are most at risk?
Older or inexpensive internet-connected products that no longer get updates, such as cameras, smart TVs, routers, and appliances, are generally the most exposed.
What is the single most useful thing I can do?
Keep everything updated and place your smart devices on a separate network from your computers and phones.
