- OpenAI revealed new AI model Sora that generates high-quality videos from text prompts.
- OpenAI is collaborating with experts on safeguards before considering wide release.
- Striking the balance between innovation and responsibility is OpenAI’s key challenge.
Artificial intelligence research company OpenAI has revealed its newest system called Sora, which can generate video content from textual descriptions and prompts.
The model is able to produce high-definition, movie-quality clips up to a minute long featuring coherent motions, characters, backgrounds, and more based on the input text.
Limitations to overcome
Selected samples displayed on Sora’s demo page showcase the model’s impressive capabilities and range—from photorealistic to animated styles.
However, OpenAI admits limitations around accurately depicting complex physics and cause-and-effect relationships. There is also a risk of confusion between spatial details like left versus right.
Responsibility at OpenAI
While positioning Sora as a research preview so far, OpenAI points to potential misuse of the technology by bad actors as a key concern needing further exploration before considering wide release.
The company says it is collaborating with experts to identify positive applications as well as build safeguards like video provenance tools.
For now, OpenAI is providing only a glimpse of Sora’s text-to-video prowess while grappling with ethical implications. However, the model demonstrates notable progress in an emergent AI field, also drawing interest from startups and big tech firms alike.
Striking the right balance between innovation and responsibility remains OpenAI’s biggest challenge going forward.
