OpenAI’s new video generator from text, Sora, is raising concerns about its potential risks of fake or misinformation and the threat to the job market.
The X, formerly known as Twitter, account of OpenAI introduced Sora, which is an AI model platform for creating realistic videos from text instructions, on Feb. 15.
According to OpenAI, Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters.
OpenAI is also the maker of ChatGPT.
OpenAI is not yet available to the users but the users of X, formerly known as Twitter, expressed concerns about the threat to the job market in the future.
William Doan, a paralegal student in Humber, is concerned about AI-generated content with potential risks like fake news or misinformation.
“I study law, so obviously, we have a lot in the context of evidence, so even AI can generate the video or any form of physical evidence content such as a video, a music piece, or a picture,” Doan said.
“If they actually can do and recognize that, it could be a big problem for us because we are not able to identify which one is the real one, which is the fake one,” he said.
“Right now, we are already having a hard time identifying AI-generated content, which is the actual one,” Doan said.
Bryan Pham, an alumnus of the Product Design program at Humber, is concerned about the copyright violation by AI to content creators like him.
“A lot of AI design comes from stealing people’s creativity. AI may have our information to create something new, but more information from artists and designers like me,” Pham said.
“That’s going to hinder a lot of creative differences, especially when trying to create something new because of the rehash of many things in society and media right now,” he said.
Pham claimed that AI could be a helpful tool for brainstorming if we use it correctly.
“Everything is just a repeat of everything else. It’s nothing original anymore,” Pham said. “It can’t be used as a crutch. I use it for inspiration; some people take ideas from others.”
OpenAI is now taking several necessary safety steps to ensure Sora is available in OpenAI’s products.
Sora is now in the testing phase by the red team with domain experts to evaluate critical areas that may be harmful and ensure that Sora is not creating misinformation or hostile or biased information.
OpenAI uploaded multiple videos created by Sora with detailed prompts, which attracted thousands to millions of interactions on Twitter.
Gokhool Muhammad Uzair Hussein Ibne Ashraf, a Business Management student at Humber, praised the evolving technology and the information AI can give.
“Technology has evolved, but AI has made it better. AI is very social because when you ask AI for something, it gives you a lot of information,” he said.
Andrei Vulcu, an experienced data analyst and programmer in the computer science field, expresses his excitement about seeing the numerous uses of data in the next few years.
“Sora can revolutionize the world of content creation by conveying a person’s idea to real life without spending millions of dollars or hundreds of working hours on animation or CGI,” Vulcu said.
“A model should act on and provide information based on its data, without bias. This depends obviously on context, the neutrality of the prompt given to the chatbot, data available, etc.” Vulcu said. “I think it’s essential that people learn to ask the right questions rather than limiting the scope of AI models.”
Besides worrying, technology users still express their excitement about the new platform.
“I have never used this kind of platform before, but if they do that one, I would really love to try it but try to eliminate the force evidence using the AI-generated as much as possible,” Doan said.