Allegedly AI-Generated Audio Bomb Threat Prompted Evacuations at the Port of Constanța, Romania

A Romanian-language audio bomb threat, assessed by officials as AI-generated, prompted employee evacuations and security searches at the Port of Constanța, Romania. Authorities received two calls using numbers with Ukrainian and Polish prefixes. Officials later said the alert was false. The perpetrators and AI system were not identified.

Risk classification

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

The incident involved malicious actors using AI-generated voice technology to conduct a targeted hoax bomb threat, manipulating the beliefs and behavior of port authorities to force an evacuation.

Additional risk subdomains

• 1.2 Exposure to toxic content: The AI system was used to generate a threatening audio message containing an illegal act and threat of violence.

Causal factors

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

The incident was caused by human actors who intentionally used an AI voice generation tool post-deployment to create and deliver a fraudulent bomb threat.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3. Limited Risk: The incident involves AI-generated audio content (synthetic voice/text-to-speech), which falls under transparency obligations for AI-generated content and deepfakes. The reports describe a 'mesaj audio in limba romana... generat cu inteligenta artificiala' (audio message in Romanian generated with artificial intelligence).

AI system and alleged parties

• AI system: None named
• AI purpose: Voice Generation
• Behaviour type: Tool
• Alleged developer: Synthetic media generation technology developers, Synthetic audio generation technology developers
• Alleged deployer: Synthetic media creators, Malicious actors, Information manipulation actors in Romania, Information manipulation actors, Hoax threat perpetrators
• Alleged harmed parties: Public safety, Ports, Port of Constanța employees, Maritime workers, Epistemic integrity

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. 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 Minor, 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 Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The reports explicitly describe the creation and delivery of a malicious AI-generated audio message containing a bomb threat.

Directly caused: An AI voice generation tool was used to directly create a malicious audio message in Romanian stating that bombs were planted in the port.

People affected

• Occurrences reported: 1
• People reportedly exposed: 100

Potential causes

Management

• Lack of AI Hoax Protocols: Port management lacked specific protocols for verifying AI-generated threats.

Technology

• AI Voice Synthesis Accessibility: Easy access to voice tools enabled creation of realistic threat audio.
• Lack of Audio Verification Tools: No systems were in place to detect synthetic or deepfake voice calls.

Data Inputs

• AI Romanian Language Training: AI models trained on Romanian enabled accurate text-to-speech output.

Human Factors

• Malicious Intent of Callers: Perpetrators intentionally used AI to automate and scale hoax threats.
• Caller Language Barrier Bypass: AI allowed foreign callers to generate fluent Romanian threat messages.

Process and Methods

• Inadequate Call Screening: Port phone systems could not filter spoofed international numbers.

Regulatory Environment

• Unregulated Voice Synthesis: Lack of strict controls on AI voice generation allows anonymous abuse.

Information quality

• Classification confidence: High
• Reason for confidence: The reports consistently describe a single, clear event: a hoax bomb threat delivered via AI-generated audio calls to the Constanta Port. The role of the AI (voice generation) is explicitly stated by official sources (the Defense Minister), and the outcomes (evacuation, false alarm) are verified across multiple news outlets.
• Ambiguities identified: There is minor ambiguity regarding the exact ministerial portfolio of Radu Miruta (referred to as interim minister, defense minister, and transport minister across different reports), but this does not affect the AI safety classification.
• Alternative interpretations: None. The event is clearly a malicious hoax utilizing synthetic audio.

Malicious actors utilized an AI-generated Romanian voice message to deliver a hoax bomb threat to the Constanta Port, leading to temporary evacuations and brief operational disruptions. The incident demonstrates how synthetic voice technology can be leveraged to increase the credibility of disruption campaigns against critical infrastructure, though the overall national security impact remains minor.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: Romania
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The immediate threat is resolved as the hoax was cleared, but the incident represents an ongoing strategic concern regarding AI-enabled harassment.
• Autonomy: Human-controlled. The AI system was used strictly as a tool (voice generator) controlled and deployed by human actors to deliver the message.
• Novelty: Evolved capability. While bomb threat hoaxes are established tactics, using AI-generated native language voice cloning to target critical infrastructure represents an evolved capability.

Impact by dimension

• Physical security: Minor. The incident targeted Constanta Port, a critical infrastructure site, leading to precautionary evacuations and temporary shutdown of utilities. However, it was a hoax with no actual kinetic threat or physical damage.
• Information security: Minor. Malicious actors used synthetic voice generation to make the hoax threat more convincing in the local language, representing a minor, localized information operation.
• Sovereignty: Minor. The threat targeted a state port authority (ANR) and triggered emergency response protocols, temporarily disrupting local government administrative functions.
• Economic security: Minor. The brief suspension of activities at Gate 1 and port administration offices caused minor operational and economic disruption, but did not threaten strategic economic security.
• Societal stability: Minor. The hoax caused localized anxiety and forced the evacuation of approximately 50 to 100 port employees, but did not trigger wider societal instability.