AI Voices Abused by Telegram User to Make Swat Calls as Paid Service

Telegram channel Torswats offered paid service for and posted own recordings of false threats calls featuring AI-generated voices to direct armed law enforcement to raid locations of victims such as high schools, private residents, streamers.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.3 Fraud, scams, and targeted manipulation

The perpetrator used synthetic voices to run a commercial swatting service, targeting specific schools and individuals for financial gain and personal advantage.

Additional risk subdomains

• 1.2 Exposure to toxic content: The synthetic voices were used to generate highly toxic, threatening, and illegal content, including fake bomb threats and mass shooting claims.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was caused by a human actor intentionally deploying synthetic voice technology to commit crimes and run a paid swatting service.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3. Limited Risk: The report describes the use of synthetic voice generation technology, which falls under the category of AI-generated content and deepfakes subject to transparency obligations.

AI system and alleged parties

• AI system: unspecified
• AI purpose: Voice Generation; Text Style Replication
• Behaviour type: Tool
• Alleged developer: unknown
• Alleged deployer: Torswats
• Alleged harmed parties: Your CBD Store, University of Pittsburgh Police Department, Phillipsburg High School, Hempstead High School, Dubuque Police Department, Bellefonte Area High School

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Substantial. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Substantial, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Minor, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes the creation and dissemination of highly malicious and threatening synthetic audio content.

Directly caused: The perpetrator generated synthetic voice recordings claiming that explosives were placed in schools or that they were going to shoot responding officers.

Inferred additional harm: At least 170 instances of highly malicious, threatening synthetic audio calls were generated and delivered to emergency services across the country.

Psychological

Reported: The report explicitly describes panic and distress resulting from the fake emergency calls and subsequent school lockdowns.

Directly caused: The synthetic calls caused immediate panic and distress to the 911 dispatchers who received the realistic threats.

Indirectly caused: Students, staff, and residents experienced severe anxiety, distress, and trauma due to emergency lockdowns (such as at Hempstead High School) and armed police responses.

Inferred additional harm: Given that up to 170 calls may have been made, it is highly likely that thousands of students, teachers, and private residents suffered acute psychological distress and trauma from these incidents.

Epistemic

Reported: The report explicitly describes the use of synthetic voices to fabricate emergency situations and deceive public safety officials.

Directly caused: The synthetic calls successfully deceived 911 operators into believing real, immediate threats existed at schools and residences.

Inferred additional harm: The systematic use of synthetic voices to fabricate emergencies across dozens of jurisdictions undermines the reliability of emergency reporting systems.

People affected

• Occurrences reported: 170
• People reportedly harmed: 2000
• People reportedly exposed: 2000

Potential causes

Management

• Platform Moderation Failure: Telegram failed to proactively block the swatting service channel.

Technology

• Synthesized Voice Tech: Allows swatters to make threats at scale and hide their true voice.
• Voice-over-Internet Tech: Enables callers to mask physical location and make untraceable calls.
• Real-time Voice Generation: Enables the synthesized voice to respond dynamically to 911 dispatchers.

Data Inputs

• Target Address Information: Addresses of schools and homes used as inputs to direct the swatting.
• Voice Actor Data: Voice recordings targeted for harassment and synthesis.

Human Factors

• Malicious Customers: Individuals like the 16-year-old ordering terrorism threats for money.
• Financial Incentives: The swatter monetized threats, charging 50 to 75 dollars per call.
• Intentional Deception: Swatters using fake narratives to trigger extreme police responses.

Process and Methods

• Cryptocurrency Payments: Allows anonymous financial transactions for swatting services.
• Telegram Distribution: Used as a platform to coordinate, advertise, and share call recordings.
• Automated Calling Scripts: Using structured scripts to systematically deceive emergency operators.

Regulatory Environment

• Jurisdictional Limitations: Swatters operate outside the jurisdiction of local law enforcement.
• Anonymity of Online Platforms: Lack of regulation preventing anonymous registration on messaging apps.

Information quality

• Classification confidence: High
• Reason for confidence: The report from Motherboard is highly detailed, providing direct evidence of the Telegram channel, specific call recordings, and official confirmation from the Dubuque Police Department regarding the use of synthetic voices.
• Ambiguities identified: The exact software or AI model used by Torswats to generate the synthetic voices is not identified.

An anonymous operator used AI voice synthesis to run a commercial swatting service, disrupting schools and targeting individuals across the US. While the incident caused localized panic and wasted emergency resources, its overall national security impact remains minor and manageable under standard law enforcement procedures.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: An anonymous Telegram user known as Torswats

Threat characteristics

• Imminence: Long-term. The threat of AI-enabled swatting represents an ongoing strategic and law enforcement concern rather than an immediate national security crisis.
• Autonomy: Human-controlled. The AI voice synthesis acted as a tool directly controlled and deployed by a human operator to conduct the calls.
• Novelty: Evolved capability. While swatting is an established threat, using real-time AI voice synthesis to scale and automate the calls represents an evolved capability.

Impact by dimension

• Physical security: Minor. The incident involved automated swatting calls that triggered emergency police responses and school lockdowns, creating a high potential for physical harm, though no actual physical injuries were reported.
• Information security: Minor. The perpetrator used sophisticated synthetic voice technology to deceive emergency dispatchers and targeted a US Representative, representing a minor information security concern.
• Sovereignty: Minor. The calls disrupted local government functions by wasting emergency response resources and forcing school lockdowns, but did not threaten core state authority.
• Economic security: Negligible. No significant threat to strategic industries, technological competitive advantage, or national financial systems was identified.
• Societal stability: Minor. The campaign caused localized panic and psychological distress among students, staff, and residents across multiple states, but did not threaten large-scale societal stability.