An Anne Arundel County Fire Department dispatcher responds to a 911 emergency call from his department’s dispatch center in Glen Burnie, Maryland, in 2020. 911 dispatchers have long dealt with crush calls, in which callers report a bomb threat, mass shooting, or some other type of life-threatening emergency with the intent of initiating a massive emergency response. Now, AI-generated voice calls pose a new challenge.
Alex Edelman/AFP via Getty Images
hide title
Alex Edelman/AFP via Getty Images
In August, a caller made a bomb threat to Wheatland Union High School in Yuba County, California.
Police responded overnight and classes were cancelled. There was no bomb and no one was injured.
Authorities said it was part of a series of hoax calls, also called swatting, made to schools across the country over the summer.
According to the FBI, Swatting calls can be made to 911 or police departments, and callers report a bomb threat, mass shooting, or some other type of life-threatening emergency with the intent of initiating a massive emergency response.
911 operators have dealt with this problem for a long time. But the incident in Yuba County was different: The call was made by an AI-generated voice, a new and growing challenge for emergency services.
“Those voices [of swatting callers] They used to be pretty easy to recognize, but now they sound natural. They can insert pauses. They may use words like “uh” or “maybe, um, let me think about it.” And it sounds very natural,” says Mel Maier, president and CEO of APCO International, the Association of Public Safety Communications Officials.
The goal of call swatting is to harass or retaliate against call recipients, such as a public official. They typically trigger a large, chaotic police response, diverting public safety resources from real emergencies.
People have also been killed and injured in crushing attacks, Maier notes.
Swatting calls made by people using AI-generated voices or background sounds make these fake calls appear even more realistic and more difficult to determine what is real and what is not.
APCO and NENA: The 9-1-1 Association, which both representing 911 dispatchers, say incidents like the one in Yuba County reflect their growing concerns about how bad actors could use AI.
APCO and the Public Safety Threat Alliance, a cybersecurity and information sharing group, investigated the prevalence of swatting in 911 call centers. They published their findings in June based on 766 responses from public safety communications professionals. About 14% of dispatchers said they were attacked using artificial intelligence tools. Maier says he believes that number will continue to rise with increasingly advanced tools.
Signs of a crushing call
Companies like ElevenLabs, an artificial intelligence company that creates synthetic voice software, offer artificial intelligence tools that are incredibly cheap and readily available, according to Brandon May, assistant professor of forensic psychology at the Florida Institute of Technology.
“I can clone my voice in ElevenLabs for less than $20 and I can change the intonations. I can change the depth or the high pitch of my voice. I can change the pauses,” he says.
And this makes it difficult for anyone to detect these AI fakes.
“It’s no longer easy to just look at an image, listen to a video, or hear a voice and say, ‘That’s AI.’ People get fooled by this all the time,” he says.
Fraudsters using AI chatbots are even tricking people into handing over their life savings, May says.
And the increasing sophistication of these fake voices and sound effects makes it especially difficult for emergency operators to deduce what is real or not during what May calls “high-risk, high-consequence” calls. “It’s going to be almost impossible to detect,” he says.
Despite the difficulty, APCO and NENA are developing ways to train their members to face these new threats.
April Heinze, NENA vice president and chief of 911 operations, says there are some characteristics of a crush call.
For example, most fly swatters do not dial 911 directly.
“The reason is that 911 provides the location. It provides a callback number. It provides the name of the person who owns the phone. It provides a lot of additional data and it’s really impossible to spoof your location on 911,” he says.
Swatting calls are typically made to police departments or courthouses, Heinze explains. These callers are eventually transferred to 911 centers. Therefore, trying to trace those calls is impossible, he says.
Training evolves to address the new threat of AI
Both Heinze and Maier decline to describe the details of training to identify potential AI crush calls. The more information that is available, they say, the easier it will be to train AI to fool dispatchers in the future.
Maier does point out an idea for training. APCO is working to create training criteria in caller interrogations, where the dispatcher asks you specific questions at “such a granular level that an AI can’t get that information from a public database, can’t capture it from a social media post, or just make a guess,” he says.
Despite growing concerns about the use of AI, both Maier and Heinze emphasize that operators will always assume that calls reporting a life-threatening situation are real and will continue to dispatch emergency resources to any call for help.
“It’s important to understand that no matter what happens, public safety will have a very strong response,” Heinze says. “Even if there are indicators that it’s a crush incident. They always report a life-or-death situation, and if it’s a life-or-death situation, you can never take any chances.”
