AI voice threats are evacuating schools: Why every phone needs AI voice detection

Earlier this week, my child’s elementary school in New York City was evacuated after a bomb threat. Hundreds of children were rushed out, terrified parents raced toward the school, and police shut down streets in Lower Manhattan. Thankfully, it was a hoax, but one phone call paralyzed an entire community. The scary truth is that a single person could launch thousands of identical threats simultaneously with agentic AI from 10,000 miles away. These threats are already illegal, but tougher laws won’t stop criminals. Technology can.

In the past, someone had to pick up a phone to deliver a threat. Today, all you need is a computer and AI.

Police haven’t said whether AI placed that call. In California this August, there was no question. Wheatland Union High School evacuated after an AI-generated bomb threat, and classes were canceled as police investigated. I’m both a parent and CEO of Reality Defender, where we build technology that detects AI-generated voices in real time, so this hits particularly close to home.

From one threatening call to thousands of AI voice attacks

For decades, law enforcement has confronted bomb threats and swatting, a tactic where callers report fake emergencies that trigger real police response. Now, agentic AI can hold convincing voice conversations, place calls, and answer questions without human involvement.

This is happening quickly. An August 2026 FBI advisory warned that swatting is shifting toward schools and flagged calls from outside the US as a warning sign. NPR reported this week that AI voices are making fake emergency calls harder to recognize. The Departments of Justice and Education have also issued guidance on swatting and anonymous threats against schools.

AI is more accessible every day, making threats nearly free to generate. This also means a single fake emergency call can mobilize expensive emergency resources. One attacker in another country can disrupt hundreds of communities at once.

More laws won’t stop criminals who already ignore them

Creating false emergency reports is already a crime punishable by fines and imprisonment. But a law only works if you can catch the people who break it. AI calling tools make perpetrators harder to identify, particularly when they’re overseas.

Passing another law against AI-generated bomb threats is like putting a “No Robberies Allowed” sign in a store window. We know robbery is already illegal, yet criminals are ignoring the rules. On their own, tougher penalties create a false sense of security. Penalties don’t stop the next call. Standards for the lines those calls travel through can.

The threat extends beyond schools. We trust voices every time we approve a payment, answer a loved one, or dispatch help. AI can now fake every one of those calls. In July, the FBI warned that scammers are using AI-generated voices to pose as people their victims know. The same technology threatens businesses, governments, and emergency dispatch centers.

At Reality Defender, we’re building the antivirus for voice communications. Our technology analyzes live call audio to detect synthetic voices in real time, within the systems organizations already use. Organizations then use this signal to manage agentic callers and apply security controls before harm occurs. It’s built to catch what the human ear can’t.

We’re working across emergency response, telecommunications, and device manufacturing to bring this protection to the platforms, systems, and devices these threats travel through. Orange Business is integrating our AI voice detection into its Live Intelligence platform, and Samsung Electronics Singapore signed an agreement with Reality Defender to deploy on-device deepfake detection to Galaxy devices. Our work with emergency responders focuses on bringing the same detection capabilities into critical communications.

Not every AI caller is malicious. AI agents offer real benefits in accessibility and customer service, but legitimate doesn’t mean cost-free. Even well-intentioned AI callers can tie up lines, repeat requests, and drain capacity that real people need. But blocking them all isn’t the answer. Detection provides a crucial security signal, letting organizations manage agentic traffic under their own policies while ensuring credible emergencies get an appropriate response.

3 changes that can’t wait

• Schools and emergency services need integrated AI voice detection.
• AI voice platforms, telecoms, and device manufacturers must embed detection as standard security infrastructure.
• Policymakers must prioritize “prevention via detection” over penalties alone, and mandate AI voice detection across critical communications.

Computers have antivirus. Email has spam filters. Phones need AI voice detection.

We cannot stop answering emergency calls, expect people to recognize synthetic voices, or assume criminals will obey the law. Real-time AI voice detection must become standard communications infrastructure. Our children and communities deserve protection as sophisticated as the threats we face.

Effective immediately, Reality Defender is offering free access to our AI voice detection technology for schools, 911 dispatch centers, and emergency responders*. Whether you’ve already received suspicious calls or want to protect your communications before an incident, our team is available 24/7 to analyze threats and deploy real-time detection. Contact Reality Defender directly to request free access and immediate support.

*Subject to program approval.