Purported Deepfake Facebook Advertisement Reportedly Used Bulgarian Actor and TV Host Mihail Bilalov's Likeness to Market Joint-Pain Product

A reportedly AI-manipulated Facebook advertisement appeared to repurpose footage of Bulgarian actor and television host Mihail Bilalov, overlaying purportedly fabricated audio and altered lip movements to market an alleged joint-pain product. The ad reportedly linked to a fabricated article imitating a Bulgarian news outlet and to an online sales funnel for a product described as unavailable through pharmacies.

Risk classification

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

The incident involves the unauthorized use of public figures' likenesses and cloned voices to deceive social media users into purchasing unverified medical products for financial gain.

Additional risk subdomains

• 3.1 False or misleading information: The deepfake videos and associated fake news articles spread fabricated information and false endorsements to mislead consumers.

Causal factors

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

The incident was caused by human malicious actors who intentionally deployed AI voice-cloning technology to create deceptive advertisements for financial gain.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3. Limited Risk: The report describes the creation and dissemination of deepfakes, which are subject to specific transparency obligations under the Act.

AI system and alleged parties

• AI system: ElevenLabs, unspecified (ElevenLabs)
• AI purpose: Voice Generation; Deepfake Video Generation
• Behaviour type: Tool
• Alleged developer: Unknown voice cloning technology developers, Unknown deepfake technology developers
• Alleged deployer: Unknown scammers
• Alleged harmed parties: People seeking medical advice, Mihail Bilalov, General public of Bulgaria, General public, 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 Minor, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: Yes, the report describes malicious deepfake videos and fake news articles created to scam users.

Directly caused: Deepfake videos mimicking Bulgarian celebrities and a fake news article mimicking Fakti.bg were created and spread via sponsored Facebook ads.

Inferred additional harm: The malicious content likely reached tens of thousands of Bulgarian Facebook users through sponsored distribution.

Epistemic

Reported: Yes, the report describes the fabrication of news articles and celebrity endorsements.

Directly caused: The creators fabricated a news article on a site mimicking Fakti.bg and generated fake audio of celebrities endorsing a product.

Inferred additional harm: This erodes public trust in online media and the authenticity of video/audio statements by public figures.

People affected

• Occurrences reported: 1
• People reportedly harmed: 6
• People reportedly exposed: 6

Potential causes

Management

• Ad Platform Profit Prioritization: Facebook accepted ad money without verifying content legitimacy.

Technology

• Deepfake Voice Generation: AI tools were used to clone the actor's voice with 98% probability.
• AI Video Manipulation: AI lip-syncing adjusted mouth movements to match the fake audio.

Data Inputs

• Hijacked TV Broadcast Footage: Original video inputs from a TV show were used without authorization.
• Unverified Audio Training Data: Public recordings of the actor's voice were used to train the AI.

Human Factors

• Targeting Vulnerable Audiences: Targeted Facebook ads reached users seeking joint pain relief.
• User Trust in Known Celebrities: Viewers believed the endorsement due to the actor's high profile.

Process and Methods

• Flawed Ad Review Process: Facebook's ad review failed to detect the AI-manipulated video.
• No Pre-Ad Verification: Lack of identity verification before running celebrity ads.

Information quality

• Classification confidence: High
• Reason for confidence: The report from Factcheck.bg is highly detailed, verifying the deepfake nature of the videos using ElevenLabs tools, identifying the specific public figures targeted, and describing the exact mechanism of the scam.
• Ambiguities identified: The exact number of users who fell victim to the scam or the total financial losses are not specified.
• Alternative interpretations: None.

A commercial scam in Bulgaria utilized AI-generated voice clones and deepfakes of local celebrities to sell unverified health products. While illustrating the growing accessibility of synthetic media for fraud, the incident has negligible direct national security implications.

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

Threat characteristics

• Imminence: Long-term. Represents an ongoing, low-level commercial scam threat rather than an active national security crisis.
• Autonomy: Human-controlled. AI tools were directly operated by human malicious actors to generate specific audio and video assets without autonomous decision-making.
• Novelty: Established threat. The use of deepfakes and voice cloning for financial fraud is an established threat methodology seen globally.

Impact by dimension

• Physical security: Negligible. No threats to physical infrastructure, kinetic targeting, or autonomous weapons deployment.
• Information security: Minor. Uses synthetic media to mislead, but serves as a commercial scam rather than a coordinated state-sponsored information warfare campaign.
• Sovereignty: Negligible. No direct threat to state authority, electoral systems, or core government decision-making processes.
• Economic security: Negligible. Limited to individual consumer fraud and did not threaten strategic industries, financial infrastructure, or national economic stability.
• Societal stability: Minor. Minor societal impact through erosion of public trust in media and digital content, alongside potential health risks from unregulated medical products.