AI Made Friendly HERE

How OpenAI Made Sora, Its Text-to-Video AI Tool – Tech News Briefing

This transcript was prepared by a transcription service. This version may not be in its final form and may be updated.

Alex Ossola: Welcome to Tech News Briefing. It’s Thursday, March 14th. I’m Alex Ossola for the Wall Street Journal.
Coming up on today’s show, some 20-somethings who spent nearly half their lives on social media have chosen to make a big move. Quit TikTok. We’ll hear why from WSJ Family and Tech columnist Julie Jargon.
And then OpenAI recently announced Sora, a new text-to-video artificial intelligence tool. In an exclusive conversation, WSJ Senior Personal Tech columnist Joanna Stern sat down with OpenAI CTO Mira Murati to better understand how the tool works and how it might change in the future.
But first, more than 170 million Americans are on TikTok, but the app’s future in the US is in question. Yesterday, the House of Representatives voted overwhelmingly to approve a bill to ban TikTok or to force the company’s sale. Lawmakers have said they fear the app’s Chinese parent company, ByteDance, would give US users’ data to China’s government.
But on a personal level, some people are already taking control of their relationship with the app. WSJ Family and Tech columnist Julie Jargon is here to tell us why some 20-somethings are calling it quits.
Julie, what is it about TikTok that makes it so hard to quit?

Julie Jargon: These 20-somethings told me that more than other social media platforms, TikTok has a particularly good algorithm. It really learns what you want to see, figures you out, and shows you content that you engage with, and it makes it hard to put it down.

Alex Ossola: Yeah, it was really notable to me, reading your story, how many people tried to quit and then ended up going back and had to quit a couple of different times.

Julie Jargon: Yeah, the pull was so strong. They felt like they were missing out on something. They missed seeing the content that they enjoyed watching.
I talked to one woman who said she would be walking her dogs while watching TikTok on her phone, and sometimes she would run into mailboxes or trees. Someone else mentioned that he was taking out the trash and he would carry his trash bag in one hand, his phone in the other, and he was worried that he was going to drop his phone in the trash bin. So it was something that was just in constant use for a lot of these people.

Alex Ossola: Based on your reporting, are we seeing a trend of younger people quitting TikTok?

Julie Jargon: There’s a mobile analytics firm that found that the number of young adult TikTok users has actually declined by about 9% between 2022 and 2023. So there has been a drop-off, and it’s hard to know yet if that is just a brief blip or if that is something that’s going to become a wider trend.

Alex Ossola: Why is it notable that this age group in particular is leaving the platform?

Julie Jargon: It’s notable because this age group was a big chunk and still is a big chunk of TikTok users. The app really exploded in popularity during the pandemic, which is when many of these 20-year-olds that I interviewed said they started really using TikTok.
What’s happening with these young people now is they have jobs now, they don’t have parents or other people telling them to go to bed and put their phones down and wake them up in the morning. And so they’re realizing that they have a lot of responsibilities and if they let their TikTok habit take over, there are going to be consequences.

Alex Ossola: Do we know if this age group leaving TikTok is in significant enough numbers that it’s affecting what TikTok sees?

Julie Jargon: Those young adults under the age of 24, there’s still a very large group of TikTok users totaling about 54 million each month, and that’s about 32% of TikTok’s 170 million monthly US users. But if that number should continue to fall, that could have an impact.

Alex Ossola: What has TikTok said about this?

Julie Jargon: TikTok says that they offer a lot of tools that can allow people to limit the time that they spend on TikTok. They have custom screen time limits that you can set on the app, sleep reminders to nudge you to get off the app and go to bed.
And the problem is these tools are only good if people use them. And the young adults I interviewed for this story said that they didn’t use those features. Some of them weren’t aware that they existed.

Alex Ossola: That was our Family and Tech columnist Julie Jargon.
Coming up, AI can now generate hyperrealistic, high-quality video. What could go wrong? That’s after the break.
When video clips generated by Sora, OpenAI’s new text-to-video model, hit social media last month, people were blown away. It rendered realistic-looking cherry blossoms. It even brought wooly mammoth seemingly to life.
OpenAI won’t publicly release Sora until later this year, but when WSJ Senior Personal Tech columnist Joanna Stern asked OpenAI to create a few new videos for her, there were some tells that they were made by AI.
She sat down with OpenAI Chief Technology Officer Mira Murati to understand more about how the model works and how the tool might get better in the future. We’re bringing you highlights from a video featuring their conversation. We’ll link the whole video in our show notes.

Joanna Stern: How does Sora work?

Mira Murati: It’s fundamentally a diffusion model, which is a type of generative model. It creates a more distilled image starting from random noise.

Joanna Stern: Okay, here are the basics. The AI model analyzed lots of videos and learned to identify objects and actions. When given a text prompt, it creates a scene by defining the timeline and adding detail to each frame. What makes this AI video special, compared to others, is how smooth and realistic it looks.

Mira Murati: If you think about filmmaking, people have to make sure that each frame continues into the next frame with this sense of consistency between objects and people, and that’s what gives you a sense of realism and a sense of presence. And if you break that between frames, then you get this disconnected sense and reality is no longer there. And so this is what Sora does really well.

Joanna Stern: You can see lots of that smoothness in the videos OpenAI generated from the prompts I provided, but you can also see flaws and glitches.
A female video producer on a sidewalk in New York City holding a high-end cinema camera. Suddenly, a robot yanks the camera out of her hand.

Mira Murati: So in this one, you can see the model doesn’t follow the prompt very closely. The robot doesn’t quite yank the camera out of her hand, but the person sort of morphs into the robot. Yeah, a lot of imperfections still.

Joanna Stern: One thing I noticed there, too, is when the cars are going by, they change colors.

Mira Murati: Yeah. So while the model is quite good at continuity, it’s not perfect. So you kind of see the yellow cab disappearing from the frame there for a while, and then it comes back in a different frame.

Joanna Stern: Would there be a way after the fact to say fix the taxi cabs in the back?

Mira Murati: Yeah. So eventually that’s what we’re trying to figure out, how to use this technology as a tool that people can edit and create with.

Joanna Stern: What data was used to train Sora?

Mira Murati: We used publicly available data and licensed data.

Joanna Stern: So videos on YouTube?

Mira Murati: I’m actually not sure about that.

Joanna Stern: Okay. Videos from Facebook? Instagram?

Mira Murati: You know, if they were publicly available, yeah, publicly available to use, there might be the data, but I’m not sure. I’m not confident about it.

Joanna Stern: What about Shutterstock? I know you guys have a deal with them.

Mira Murati: I’m just not going to go into the details of the data that was used, but it was publicly available or licensed data.

Joanna Stern: After the interview, Murati confirmed that the licensed data does include content from Shutterstock.
Right now, Sora is going through red teaming, aka, the process where people test the tool to make sure it’s safe, secure, and reliable. The goal is to identify vulnerabilities, biases, and other harmful issues.
What are things that just you won’t be able to generate with this?

Mira Murati: Well, we haven’t made those decisions yet, but I think there will be consistency on our platform. So similarly to DALL-E where you can’t generate images of public figures, I expect that we’ll have a similar policy for Sora. And right now we’re in discovery mode and we haven’t figured out exactly where all the limitations are and how we’ll navigate our way around them.

Joanna Stern: What about nudity?

Mira Murati: I’m not sure. You can imagine that there are creative settings in which artists might want to have more control over that. And right now we are working with artists and creators from different fields to figure out exactly what’s useful, what level of flexibility should the tool provide.

Joanna Stern: How do you make sure the people who are testing these products aren’t being inundated with illicit or harmful content?

Mira Murati: That’s certainly difficult, and in the very early stages it is part of red teaming, something that you have to take into account and make sure that people are willing and able to do it. When we work with contractors, we go much further into that process, but that is certainly something difficult.

Joanna Stern: We’re laughing at some of these videos right now, but people in the video industry may not be laughing in a few years when this type of technology is impacting their jobs.

Mira Murati: You know, the way that I see it is this is a tool for extending creativity and we want people in the film industry, creators everywhere, to be a part of informing how we develop it further and also how we deploy it, and also what are the economics around using these models when people are contributing data and such.

Alex Ossola: That was WSJ Senior Personal Tech columnist Joanna Stern speaking with OpenAI CTO Mira Murati. You can check out the whole video of their conversation linked in our show notes.
And that’s it for Tech News Briefing. Today’s show was produced by Julie Chang with supervising producer Katherine Milsop. I’m Alex Ossola for the Wall Street Journal. We’ll be back this afternoon with TNB Tech Minute. Thanks for listening.

Originally Appeared Here

You May Also Like

About the Author:

Early Bird