Last week, Meta hit hard with l’annonce de son nouveau modèle d’IA, Movie Gen! A tool capable of generating videos with sound from simple text descriptions or images. This announcement captured the attention of AI specialists on social networks, putting Meta at the centre of discussions. Can this technology really transform the entertainment industry? And what are the ethical issues surrounding this technology?
A new AI that’s shaking up the creative industry
Movie Gen not only generates videos in a matter of seconds, it can also add sound effects, music and sound effects, all controlled by text requests.
Meta presented some impressive examples, including a video of penguins dressed in Victorian garb and a quad bike driving through a desert with a roaring engine. These demonstrations left internet users speechless, giving the feeling that Meta may (finally?) have a major trump card in the generative AI race.
I couldn’t be more excited to share our latest AI research breakthrough. We call it Meta Movie Gen and it’s a collection of state-of-the-art models that combine to deliver the most advanced video generation capability ever created.
Check it out: https://t.co/K4IxAVPO3y pic.twitter.com/HawreKeavy
— Ahmad Al-Dahle (@Ahmad_Al_Dahle) October 4, 2024
This is wild.
Meta AI just announced Movie Gen that generate AI video just from text.
And it can generate sound effects and music for your videos.
100% AI (including sound) 🔊
1. Penguins swimming w/ a beautiful orchestral piece that evokes a sense of wonder. pic.twitter.com/y29Q46k3rB
— Min Choi (@minchoi) October 4, 2024
This could change video forever.
Meta just announced @meta Movie Gen, their new AI video model that can do text to video, video to video, and text to audio. pic.twitter.com/6C32uFegGT
— Ravi Vora (@RaviVora) October 4, 2024
One of the most innovative aspects of Movie Gen is the ability to transform a simple image or video into a complex production, with virtually unlimited modifications. You can change the background, objects, add visual and sound effects – all simply by describing what you want.
Summary of the text used to generate the video: A female DJ spins records on a rooftop in Los Angeles. She is wearing a pink jacket and giant headphones. A cheetah is standing next to the woman. The background is a cityscape.
Fierce competition but a unique sonic asset for now
Now HEAR this (not just watch) – We’ve got audio covered for generated videos 🔊
Introducing Movie Gen Audio, which adds 48kHz synced SFX and aligned music to amazing videos from Movie Gen Video (and other sources!)
Super honored to work with this amazing team!
More to come🔥🔥 https://t.co/TnQ7pvS9uZ— Wei-Ning Hsu (@mhnt1580) October 4, 2024
While Meta is not the first company to venture into generative video, Movie Gen stands out for its stunning realism and its ability to integrate audio. Where competitors such as Sora d’OpenAI or Gen-3 Alpha de Runway focus on visual creation, Meta goes further by offering videos with a full soundtrack, providing a real competitive edge in video production… enough to worry Hollywood.
The past year has been marked by intense debate within the audiovisual sector. While technophiles hail generative AI as a lever for speeding up production and pushing back the boundaries of creation, others are concerned about the potential abuses. The use of protected works without authorisation, the threat of actor replacement and the proliferation of synthetic content raise complex ethical issues that loom large over this technological revolution.
In 2023, the industry had already been shaken by a screenwriters’ strike, linked to the growing use of artificial intelligence in scriptwriting, provoking an unprecedented mobilisation.
Movie Gen is not yet available to the general public and remains in the testing phase. As with its predecessors, it will probably be some time before it is deployed on a large scale. Meta has not yet announced an official launch date.
The age of permanent doubt: the risk of unidentifiable deepfakes
Beyond the creative world, this spectacular innovation raises ethical questions for many other sectors. By making video creation so simple and accessible, Movie Gen could also lead to an explosion of manipulated content in the form of deepfakes.
A simple snapshot could be used to create compromising or politically dangerous photos and videos. This is a particularly worrying issue at a time of major elections, when social networks are increasingly becoming citizens’ priority source of information. It is also a new potential source of destabilisation in geopolitical relations, with an amplification of the propaganda of states, separatist groups and protest movements.
In everyday life, creating false identities for dating sites has never been easier. People will be able to invent lives, families and friends in just a few clicks.
While Movie Gen opens up infinite creative possibilities, it also raises worrying ethical questions. The ease with which we can now create hyper-realistic videos threatens to destabilise our very relationship with reality.
Summary of the text used to generate the video: Make a nice selfie video of a man and his dog. The man is wearing a black T-shirt. The dog is a beagle puppy. The background is a patio full of trees. The man has a big smile on his face as he tries to take the perfect selfie with his dog. The lighting is warm.
By Pierre Bellagamba