WPP CEO Mark Read targeted by AI voice cloning and deepfake scam
Scammers attempted to defraud WPP executives by using AI-generated voice and video to impersonate CEO Mark Read. The attack involved a fake WhatsApp account and a Microsoft Teams call but was ultimately unsuccessful.
Criminals used AI voice cloning and public footage to impersonate a CEO in a real-time virtual meeting.
Key facts
- What
- Scammers attempted to defraud WPP executives by using AI-generated voice and video to impersonate CEO Mark Read.
- Incident date
- May 10, 2024
- Who
- WPP
- Failure mode
- Hallucination
- AI surface
- Voice Agent
- Severity
- Low
What happened
Fraudsters impersonated WPP CEO Mark Read using AI-generated voice cloning and YouTube footage to deceive company executives. The attackers used a fake WhatsApp account to coordinate a Microsoft Teams meeting to facilitate the scam. The attempt was unsuccessful and did not result in financial loss.
What broke inside the model
- 01 · TriggerA user asks for a fact, a citation, or a figure.
- 02 · Model stepThe model writes a fluent, confident answer.
- 03 · Control gapNothing ties the claim back to a real source.
- 04 · FailureA fabricated fact ships as if it were verified.
- 05 · ConsequenceThe false claim reaches a customer, a court, or the public.
Confidence holds, and even spikes, as the claim detaches from any source.
The authentication mechanism for virtual meetings failed to distinguish between a live human and an AI-generated deepfake. Attackers exploited the vulnerability of visual and auditory identity verification by using generative AI to synthesize a convincing impersonation of the CEO.
What it cost
Sources
Cite this entry
https://failureindex.ai/failures/wpp-ceo-mark-read-targeted-voiceAI Failure Index. "WPP CEO Mark Read targeted by AI voice cloning and deepfake scam" (FI-0495). Realm Labs. https://failureindex.ai/failures/wpp-ceo-mark-read-targeted-voice (indexed Jun 10, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0495. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
- AI Detection & Response (AIDR)
A runtime layer that watches the model's internal state can flag the moment a model commits to a claim it has no support for, and hold or reroute the response before it reaches a user. Realm reads those signals in real time rather than grading the transcript after the fact.