With the launch of its latest product, the creator of ChatGPT has promised to revolutionise video production. Can it live up to the hype?
ChatGPT for Video
On 15 February, OpenAI, the company behind the ChatGPT large language model, announced new video generating software. Named after the Japanese word for sky, Sora has been opened up to a small group of researchers and video creators, while examples of videos produced using Sora have been released. Producing videos up to sixty seconds long, Sora can apparently animate still images, connect videos together, extend them beyond their current limits, and create entirely new moving images based on written text. If it lives up to the hype, it will provide a powerful new tool for video creators. And given the prevalence of TikTok, now is the perfect time for videos limited to a minute or less.
“While the potential of tools like Sora is undoubtedly spectacular, it raises serious questions for users and investors.”
Facing AI’s Limitations
Like any new piece of software, what’s on show reflects Sora’s limitations and its strengths. Early commentators have seen some strange video effects, but it seems to perform better than similar video software. Clearly, this is going to get better over time and testing, creating smoother results, but another piece of news raised questions about how much better. Sora was announced less than a week before ChatGPT, OpenAI’s most prominent product, suffered a public meltdown, providing users with strange and unusable responses. That’s not the only worrying sign for OpenAI. The company is currently facing multiple court cases over ChatGPT, and last year saw coup and counter-coup in the boardroom. Serious questions are hanging over both the ethics and the long-term viability of parts of its business.
What Does Sora Mean for the Media Industry?
AI tools are already starting to have an impact in this sector, and Sora will add to this. If it works as promised, then it could save huge amounts of time on video production, cutting out repetitive and fiddly work. But as similar software for pictures and text have shown, there are limitations on the results. ChatGPT is notorious for its generic and error-filled outputs, which provide filler text on a limited budget but don’t meet the standards of professional media companies. While some creators are learning and adapting to new tech, others are resisting its encroachment, causing conflict with companies and employees. And while AI’s novelty has attracted venture capitalists, it’s not clear how profitable it can be at scale, given the huge amounts of hardware and energy involved. Will these tools actually be more cost effective than human video makers? While the potential of tools like Sora is undoubtedly spectacular, it raises serious questions for users and investors.
