This is the latter part of a 2-part series on Sora. Both parts have been released at the same time. If you haven’t already, please listen to part one first. 
As for the safety aspect, on top of working with red teamers to test the AI model, the company is reportedly developing tools to detect misleading content. OpenAI cites a “detection classifier that can tell when a video was generated by Sora” as an example of such tools that are in development. They're also using the safety methods they are using for their existing products such as DALL·E 3. 
The company has also announced that they would like to incorporate C2PA metadata if and when Sora is added to an OpenAI product. According to OprnAI, “our text classifier will check and reject text input prompts that are in violation of our usage policies, like those that request extreme violence, sexual content, hateful imagery, celebrity likeness, or the IP of others. We’ve also developed robust image classifiers that are used to review the frames of every video generated to help ensure that it adheres to our usage policies, before it’s shown to the user.”
Note that the above describes what would happen if the model is incorporated into an OpenAI product, which has not yet happened and probably will not happen anytime soon, considering Sora was only announced recently. 
Of course, their announcement of the product came with a disclaimer: “Despite extensive research and testing, we cannot predict all of the beneficial ways people will use our technology, nor all the ways people will abuse it.”
That's it for today, but if you are interested, you can learn more about the technology by reading the technical report OpenAI has published. 
Thank you for listening to The Old and the New and see you next episode.