Sora Means Text To Video Has Officially Arrived

Artificial intelligence has become one of the biggest talking points in recent years, and for good reason. It’s come so far in such a short space of time, prompting plenty of concerns that it’s developing at too fast a rate. If you need proof of how far it’s come already, you only need to look at what OpenAI has revealed recently as the company continues to make waves in this industry.

In mid-February 2024, they showed off a new tool that allows you to create videos from text. Named Sora, the technology is able to produce clips ranging up to a minute in length in a variety of ways, which naturally has its ups and downs. On the one hand, it can give people more freedom in the things they create and potentially help with roadblocks in certain projects. On the other, though, it risks making it even easier for those who use AI for nefarious means to create damaging or misleading content. So, should we be concerned?

Sora can create video from text, still images, and existing footage

It was inevitable that the development of AI technology would eventually lead to something like Sora. Once artificial intelligence was able to create text and imagery that was believably authentic, video always seemed like the next logical step. OpenAI’s new tool may well be more advanced than people were expecting at this stage in development, though. Apparently, not only can it produce a video from text, but it’s also able to create a clip from a still image or extend an existing video. You can certainly see how this can prove useful. For instance, someone doing a film project who forgot to take a particular shot or isn’t able to do so because of external factors could use something like Sora to fix their problem. However, not everyone would want to utilize it for such innocent means.

Efforts are reportedly being made to ensure the safety of Sora

Thankfully, when announcing Sora, OpenAI also revealed that access is only available to a specific collection of people right now. These researchers and creators are testing the tool to ensure that people aren’t able to exploit it and use it in ways that go against the terms of service. This includes creating footage of extreme violence, hateful imagery, celebrity likenesses, and other inappropriate content. It’s essential that Sora not be able to be exploited in such a way before it becomes available to the wider world. Otherwise, you risk the internet becoming full of videos that could damage careers, influence elections, and so much more. While the efforts of these researchers and creators don’t mean things won’t slip through the cracks, they should ensure that Sora is a somewhat safe tool to release to the public.

Sora’s arrival will undoubtedly prompt competition to start cropping up

Sora is not the first of its kind, with other AI companies releasing their own versions of video-generating AI in recent years. However, none of them have accomplished what OpenAI seemingly has, with their videos much shorter and less accurate to the prompts. Naturally, now that we know about Sora, other companies will likely increase their efforts to try and compete with this tool. Google and Meta have already shared plans for their own version of Sora, although we’ve not seen what their efforts have led to yet. We don’t doubt that by the end of 2024, the AI landscape will look a lot different from how it does now, thanks to such developments. Whether or not that’s a good thing, though, largely depends on how you view artificial intelligence as a whole.

OpenAI hasn’t always attracted the best publicity, what with the various lawsuits accusing the company of copyright infringement. Supposedly, the footage used to train Sora doesn’t breach any copyright, although this is something we expect people will be looking at quite closely. After all, with the potential damage that AI can do, we can’t afford to be too careful now that text-to-video content has become a reality.