OpenAI’s video generation platform, Sora, is facing criticism over concerns that it reproduces and perpetuates biased representations of gender, race, and physical ability. The tool, which transforms text prompts into realistic video clips, has been found to reflect outdated and stereotypical portrayals of people in a range of professional and social roles.
Observers note that Sora tends to depict roles like pilots, CEOs, and professors as male, while presenting flight attendants, receptionists, and childcare workers as female. These patterns suggest a reliance on long-standing gender stereotypes. Similarly, the portrayal of disabled individuals appears to lack diversity, with most representations limited to people using wheelchairs. Racial diversity has also been called into question, as the tool frequently defaults to youthful, conventionally attractive, and able-bodied characters, often overlooking broader demographic realities.
Industry experts warn that if such biases go unaddressed, AI-generated content could further entrench harmful narratives and present an inaccurate picture of society. They stress the need for AI developers to train models using more inclusive and representative datasets, while also conducting thorough evaluations of the broader social impact these technologies may have.
As artificial intelligence becomes increasingly embedded in media and communications, ensuring fairness and accuracy in AI-generated representations is becoming a major challenge. The responsibility now lies with tech developers and regulators to ensure these tools reflect the full spectrum of human experience.
Stay tuned to DC Brief for further updates on this story and other technology developments.
