Discussion

AI voice hoaxes are making fake 911 emergencies harder to spot

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#12603

AI-generated voices are making swatting calls harder for emergency dispatchers to recognise, NPR reports. A survey found that about 14 per cent of 766 public-safety communications professionals said they had been targeted using AI tools; dispatchers still have to treat reports of life-threatening emergencies as real.

Watch Desk analysis

What happened

NPR describes an August bomb-threat call to a California high school that authorities said was part of a series of swatting hoaxes. The call used an AI-generated voice. Police responded overnight, classes were cancelled, and no bomb was found or anyone hurt.

The calls aim to trigger a major emergency response through false reports of threats such as bombings or shootings. Mel Maier, head of APCO International, told NPR that synthetic voices can sound natural, including pauses and hesitations. APCO and the Public Safety Threat Alliance published survey findings in June, based on 766 responses from public-safety communications professionals. About 14 per cent said they had been targeted using AI tools.

Why it matters

A convincing fake can still bring police rushing to a school or other target, diverting resources from genuine emergencies and putting people at risk. The challenge is not simply spotting a synthetic voice: dispatchers cannot safely dismiss a call because it sounds suspicious.

NENA and APCO are developing training for their members. NENA’s April Heinze told NPR that dispatchers will continue to send help when a call reports a life-or-death situation. That leaves responders trying to assess credibility without delaying the response a real emergency needs.

Our read

This is a difficult security problem with no neat “AI detector” button. NPR reports that APCO is developing training around detailed questions that may be harder to answer using information scraped from public sources. The goal is to improve assessment without giving potential callers a public cheat sheet. Sensible, if not especially tidy.

What to watch

  • Whether APCO and NENA publish practical training guidance for dispatchers.
  • Whether further survey findings show how the reported use of AI in swatting is changing.
  • How emergency services can assess suspicious calls without slowing responses to genuine threats.

Discussion spark: Should emergency services share more about how they identify AI-assisted swatting, even if that could help callers evade detection, or keep the methods largely private?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.