Advocacy group calls on OpenAI to address Sora 2’s deepfake risks

This report examines growing misuse risks from OpenAI’s Sora 2 video generation model, noting that improved realism and easily removed watermarks enable convincing deepfakes that can amplify disinformation, harassment, and fraud. Advocacy groups urge a pause and stronger guardrails, while OpenAI cites watermarks, provenance metadata, and opt-in likeness controls but acknowledges detection limits. Examples include political misinformation, racist depictions of deceased public figures, and concerns about sexualized content, suggesting existing safeguards may be insufficient.