Meta Withholds Release of Powerful Voicebox AI Model Due to Misuse Risks

Thumbnail Image

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

Meta developed Voicebox, an advanced AI speech generation model capable of high-quality, versatile audio tasks. Citing concerns over potential misuse for disinformation or impersonation, Meta decided not to release the model or code, sharing only audio samples and research papers to mitigate risks of harm.[AI generated]

Why's our monitor labelling this an incident or hazard?

The article explicitly mentions an AI system (Voicebox) capable of generating realistic speech and highlights the potential for misuse leading to harms like fraud or fake videos. However, since the system has not been released or used in a way that has caused actual harm, this constitutes a plausible risk rather than a realized incident. Therefore, this event fits the definition of an AI Hazard, as the development and potential use of the AI system could plausibly lead to harm, but no harm has yet occurred.[AI generated]
AI principles
AccountabilityPrivacy & data governanceRespect of human rightsRobustness & digital securitySafetyTransparency & explainabilityDemocracy & human autonomy

Industries
Media, social platforms, and marketingDigital securityGovernment, security, and defenceConsumer services

Affected stakeholders
General public

Harm types
ReputationalPublic interestHuman or fundamental rightsEconomic/Property

Business function:
Research and development

AI system task:
Content generation


Articles about this incident or hazard