I think we can all acknowledge there are lot of drawbacks to social media. From misinformation to body shaming, social media can cause a lot of harm. However, I encourage you to take a moment and think about the best parts of social media. You can reconnect with people who you don’t see as much anymore, learn new skills and interact with people who share your sense of humor or style. 

So, with your favorite things about social media in mind, how about we take all that and get rid of it?  Namely, the word “social.” Instead, let’s just replace all the videos of your friends, family and acquaintances with an Artificial Intelligence (AI)-generated video of your stuffed animal dancing. 

Well, that’s Sora, OpenAI’s video-generation app.

Sora debuted last year, but its more recent version was released in October. The new Sora update allows you to incorporate yourself into AI videos, through uploading images and videos of your face. 

However, these AI-generated videos aren’t your typical AI slop. Even those who pride themselves on their ability to determine whether content was made with AI will struggle with the technology that Sora is piloting. These videos are strikingly realistic, blurring the lines between fiction and reality.

So now we come to the issues that are so common with discussions about AI: copyright and safety. Although OpenAI claims there are safeguards around copyrighted content, copyrighted characters and illegal content can still be found on Sora. 

And despite Sora’s restrictions concerning violent and inappropriate content, NPR’s tests of the app revealed that it was possible to create videos confirming conspiracy theories or videos that broke anti-violence content policies.  

So what are the consequences of very real-looking, dangerous videos floating around on the internet?

Yes, I’m aware that deepfakes and AI-generated videos existed well before Sora, but the ability that Sora demonstrates is what makes it so frightening.

In a world where false information is already amplified by social media, the implications of this technology are terrifying. Advanced AI-generated video can be weaponized not only to spread misinformation, but to create content that harms specific people and is weaponized against public figures in cruel ways. 

In mid-October, OpenAI decided to block users from making videos of Martin Luther King Jr. after his estate brought up their concerns about deepfake racist videos of King that were circulating on the app. 

Although Sora ultimately promised control of the likenesses of historical figures to their estates, this is indicative of a larger trend. Sora appears to be playing the same game as so many other companies in the AI race. Namely, “let’s leave this as open as possible until people are hurt or angry and we have to promise to be better to avoid getting sued.” Not my most succinct description, but it is accurate. 

This technology is new and powerful and has a concerning lack of content restrictions, or at least content restrictions that are seemingly easy for users to bypass. Questions about harmful content and copyright violation in regards to AI aren’t going away. So, in a world where an increasing number of Americans are more concerned than excited about AI, Sora only adds insult to injury. 

The views in this column do not necessarily reflect the views of the HiLite staff. Reach Charlotte Moser at [email protected].