Microsoft is working on artificial intelligence that turns photos into videos of “talking faces”

“L“The goal is not to create content intended to mislead or deceive,” the company clarifies, but acknowledges that, like “other methods of content creation, it can be misused to impersonate a person.” .

The rapid growth of generative artificial intelligence, which makes it easy to create all kinds of content (text, images, sounds, etc.) of amazing quality, has raised many concerns, especially in terms of being used for fraud and disinformation.

“Our research is focused on (…) virtual avatars with a view to their positive applications,” says Microsoft. “We oppose any behavior that aims to create misleading or harmful content from real people.”

Therefore, the IT company, the main investor of OpenAI (ChatGPT), does not plan to provide a new tool or technical information “until we are confident that the technology will be used responsibly and in accordance with current regulations.”

A computer program called VASA-1 takes a simple portrait photo and audio file and converts them into a video where the face comes to life and speaks in a hyper-realistic manner.

In the online examples – previews made from images of virtual characters – the lip movements are synchronized, the facial features are convincing, and the head movements are a little jerky, but still impressive.

Other companies are working on this technology, such as Runway, one of the specialists in generative artificial intelligence for video. Google researchers also created a Vlogger artificial intelligence model that can generate realistic videos of talking heads.

According to Microsoft, the benefits of developing this technology, “such as increasing equity in education, improving accessibility for people with communication difficulties, providing therapeutic support to people who need it, and others,” justify this research.

Many authorities are considering regulating generative AI. In December, the European Union agreed to unprecedented legislation to regulate AI, aiming to encourage innovation while limiting potential abuse.