BBC presenter’s AI-generated voice used to trick company
Occurred: March 2024
Report incident 🔥 | Improve page 💁 | Access database 🔢
A BBC presenter's image was used in an ad campaign without her consent after an AI deepfake of her voice was used by a scammer to negotiate the deal.
Insect repellant company Incognito was tricked into believing it was speaking to BBC science presenter, Liz Bonnin, through a series of voice messages and a Facebook profile allegedly purporting Bonnin's identity. The deal was in fact negotiated between a scammer and Incognito CEO Howard Carter via email and Whatsapp voice messages. Incognito paid the scammer £20,000, believing it to be Bonnin.
AI experts analysed the voice messages and agreed the voice was likely to have been generated using AI given the inconsistency in accent and the voice’s cadence and monotony.
The incident was seen as a warning about potential misuse of AI voice replication technologies In addition, Incognito lost money and faced reputational damage.
System 🤖
Unknown
Operator:
Developer:
Country: UK
Sector: Media/entertainment/sports/arts
Purpose: Generate voice
Technology: Deepfake - audio
Issue: Personality rights; Fraud
Transparency: Governance
Page info
Type: Incident
Published: May 2024