Skip to content
Submit incident
AII-1002-58657
ReportedReported by fewer than three independent publishers, with no official finding.

Brian Cranston voice and likeness used without consent in Sora training

Breaking Bad actor Bryan Cranston's voice and facial likeness were used without his permission to train OpenAI's Sora video generation model. After Cranston and actors' union SAG-AFTRA raised concerns, OpenAI strengthened safeguards and committed to opt-in protocols for protecting performers' identities.

October 2025 · OpenAI · Intellectual property and appropriation
October 2025
AII-1002-58657View source ↗
Quick facts
Category
Intellectual property and appropriation
Country
United States
Source
sagaftra.org
Date
October 2025
Drafted by
Raidu Incident Detection

What happened

Unauthorised AI-generated clips using Cranston's voice and likeness appeared on the Sora app, including a video appearing to show late pop star Michael Jackson taking a selfie with Walter White, Cranston's character from Breaking Bad.

OpenAI relied on large datasets that included publicly available media, and had few safeguards in place to prevent the system from producing unauthorised imitations of celebrities and copyrighted content.

Reported impact

For actors, the case highlights the loss of control over their own image and voice in the era of generative video, and raises significant concerns over consent, intellectual property rights, and the actual or potential misuse of their identities.

It fuels ongoing industry anxiety about AI's role in replacing human talent and undermining creative livelihoods.

Index analysis

OpenAI claimed the Cranston incident was an unintentional generation, but is widely seen as having deliberately ignored the interests of rightsholders when it launched Sora 2 in a bid to get the app to market quickly, increase awareness, and make money.

The company's initial policies did not fully prevent the use of people's likenesses or voices unless they explicitly participated in the platform's cameos feature, which requires permission.

Governance controls

After talks with actors' union SAG-AFTRA and Cranston, OpenAI said it would strengthen guardrails around unapproved AI-generated content and stated it is committed to responding expeditiously to any complaints.

OpenAI updated its policy to allow rightsholders more granular control over generation of characters and committed to using an opt-in protocol where all artists have the ability to control replication of their voice and likeness.

Classification

Category
Intellectual property and appropriation
Industry
Arts, entertainment and recreation
Country
United States
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Reputational
Harmed
Workers
Failure modes
Training data misuse, Deepfake video, Voice clone, Content moderation failure
AI systems
Sora

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

sagaftra.org
Brian Cranston voice and likeness used without consent in Sora training
Oct 2026Carlsen Verlag sues OpenAI for copyright infringement of "Das NEINhorn"AllegedIntellectual property and appropriationSpiegel OnlineRP OnlineBörsenblatt+5Oct 2026Argentine media group Perfil sues OpenAI and Microsoft over unauthorized content useAllegedIntellectual property and appropriationPerfilClarinSep 2026Mathematician Accuses OpenAI of Using Unpublished Research in Model TrainingAllegedIntellectual property and appropriationThe VergeCybernewsQuartz+2Sep 2026OpenAI's GPT-6 Astra shows academic integrity concerns in mathematics researchAllegedIntellectual property and appropriationĐọc báo tin tức, tin mới Ngày nay OnlineVietnamPlusAug 2026AI companies buy and destroy rare books for model trainingAllegedIntellectual property and appropriationThe Wall Street JournalOman Observer