The new OpenAi social network would be Tiktok if it was a flow of soil ai

Welcome to the era of anti -social media. According to a Wired report, Openai plans to launch an autonomous application for its Sora 2 video generation tool which will include a Tiktok style video scrolling that will allow people to scroll through the videos entirely generated by AI. The quixotic effort follows the recent Meta launch of a flow only of the AI-Slop on its Meta AI application which has encountered almost universal negativity.

According to Wired, the Sora 2 application will include familiar style navigation in Slip-up-to-Tt-to which is presented for most vertical video platforms like Tiktok, Instagram Reels or YouTube Shorts. He will also use a personalized recommendation algorithm to feed the content of users who could use their interests. Users will be able to like, comment or “remix” an article – all very standard social media dishes.

The big difference is that all the content of the platform will be generated by AIA via the OpenAi videos generation model that can take text, photos or videos existing and Ai-Si-IFIFY. The videos will last up to 10 seconds, probably because it is about how long Sora can stand together before starting to hallucinate the strange shit. (The first version of Sora allows videos up to 60 seconds, but has trouble producing really convincing and continuous images for so long.) According to Wired, there is no way to directly download a photo or video and publish it not published.

Interestingly, Openai understood how to work a social element in the application, although in a way that has a kind of inherent goose flesh. By wire, the Sora 2 application will ask users to check their identity via facial recognition to confirm their resemblance. After confirming their identity, their resemblance can be used in videos. Not only can they fit into a video, but other users can tag and use your resemblance in their videos. Users are informed whenever their resemblance is used, even if the video generated is saved in drafts and never published.

How it will be implemented when and if the application launches to the public, we will have to see. But as indicated, it seems to be an absolute nightmare. Basically, the only thing the federal government has managed to find a kind of consensus with regard to AI regulation is to provide limited protections against non -consensual depths. As described, this kind of operation resembles a characteristic of Sora 2 is to let your resemblance be manipulated by others. There will surely be a kind of deactivation available or the ability to restrict that can use your resemblance, right?

According to Wired, there will be protections regarding the type of content that Sora 2 will allow users to create. He is trained to refuse to violate copyright, for example, and would have filters in place to restrict certain types of videos to be produced. But will he really offer sufficient protection to people? Openai has made a big point to emphasize how he added protections to the original Sora model to prevent him from generating nudity and explicit images, but the system tests have managed to have it created anyway, prohibited content at a low-zero rate.

Gizmodo contacted Openai to confirm his plans for the application, but did not receive an answer at the time of publication. There has been speculation for months on the launch of Sora 2, in the hope that it would be announced at the same time as GPT-5. For the moment, she and her support application remain theoretical, but there is at least a good idea hidden in the concept of social food entirely in AI, although probably not in the way in which the intention operates: keep the content of AI in quarantine.