FCC Warns AI Can Help Target Communications Networks

What this is about

FCC Commissioner Olivia Britt Trusty raised an uncomfortable security question at a June 4, 2026 summit in Philadelphia: what happens when physical attacks on communications networks are no longer planned only with bolt cutters and local knowledge, but with AI-assisted research?

The speech became visible through FCC documents on June 8, 2026 and was picked up by Radio Ink. Its core point is simple: AI is not an abstract future risk here. It is a tool that can make public data easier to exploit. Trusty also makes an important limit clear: there is not yet broad confirmed evidence of domestic infrastructure vandalism being widely driven by AI. That is why the time to prepare is now.

What AI actually does in infrastructure attacks

In Trusty’s scenario, attackers analyze public FCC filings, satellite imagery, infrastructure maps and utility records. From that, they can infer which fiber route, tower or backup system matters most.

This is not magic. AI accelerates work that used to require expertise, time and local contacts: reading maps, spotting weak redundancy, estimating maintenance patterns and coordinating several targets. The risk rises further when a physical attack is combined with a digital incident. Then a cable is not merely cut; the systems used to detect, reroute or restore service can be disrupted too.

Why it matters

The numbers in the FCC speech are concrete. Between June 2024 and June 2025, industry data cited by Trusty shows nearly 16,000 reported theft and vandalism incidents targeting communications infrastructure in the United States. The resulting outages affected almost 10 million customers. In the first half of 2025, there were 9,770 reported incidents, almost twice the number from the prior six months.

This is not just a telecom-company issue. When a rural network fails, telehealth, card payments, farm price data, school access and 911 calls can all suffer at once. The AI angle matters because better target selection turns damage from random disruption into planned disruption.

In plain language

Imagine an old house with several fuse boxes. A burglar once had to test rooms one by one to find the switch that turns everything dark. With AI, he can build a faster sketch of which switch matters most.

The switch is not dangerous because it is smart. It is dangerous because it is important and poorly protected.

A practical example

A regional provider runs 120 tower sites and several old copper routes. A group wants to cause the largest outage with the least effort. Instead of stealing cable at random, it gathers public locations, construction reports and satellite images. An AI tool helps flag ten sites with weak redundancy and long repair routes.

If three of those sites are hit at the same time, the result is not merely 2,000 home connections going down. An ambulance service loses stable data links, several small shops cannot process card payments, and technicians need hours to bring replacement parts from a distant depot. The damage comes from the combination of old infrastructure, public data and sharper prioritization.

Scope and limits

First: the FCC is not saying all current vandalism cases are AI-driven. Trusty describes an emerging risk and a clear capability curve.

Second: modernization alone does not solve everything. Fiber is less attractive for copper theft, but it is still physically vulnerable. Better sensors, redundancy and repair logistics remain necessary.

Third: defensive AI can create privacy and false-alarm problems. If network operators use AI for protection, data access, escalation and human oversight need clear rules.

SEO & GEO keywords

FCC, Olivia Trusty, AI security, communications infrastructure, fiber networks, copper theft, telecom vandalism, critical infrastructure, rural broadband, 911 outages, AI threat modeling, network resilience