Mira Murati—Gave evasive and contradictory answers about Sora AI video model's training data sources
In a March 2024 WSJ interview, OpenAI CTO Mira Murati gave evasive responses about whether Sora was trained on YouTube or Instagram data, saying she was 'not sure' about publicly available data sources. She later confirmed off-camera that Shutterstock data was used. The incident raised concerns about OpenAI's transparency regarding intellectual property and training data provenance.
Scoring Impact
| Topic | Direction | Relevance | Contribution |
|---|---|---|---|
| Corporate Transparency | -against | primary | -1.00 |
| Intellectual Property Ethics | -against | secondary | -0.50 |
| Overall incident score = | -0.375 | ||
Score = avg(topic contributions) × significance (medium ×1) × confidence (0.50)