MY KOLKATA EDUGRAPH
ADVERTISEMENT
regular-article-logo Saturday, 06 July 2024

Tribeca Film Festival to feature shorts using artificial intelligence

The festival will screen five short films made using OpenAI’s Sora, a text-to-video artificial intelligence model that generates video clips based on written descriptions

Mathures Paul Calcutta Published 06.06.24, 10:48 AM
Representational image

Representational image File image

The Tribeca Film Festival (TFF), founded by Robert De Niro, Jane Rosenthal and Craig Hatkoff, is returning this year to New York City to debut films from a unique source: OpenAI, the company behind ChatGPT.

The festival will screen five short films made using OpenAI’s Sora, a text-to-video artificial intelligence model that generates video clips based on written descriptions. The films, to be shown as part of Sora Shorts on June 15 (the festival will run from June 6 to 16), are by Bonnie Discepolo, Ellie Foumbi, Nikyatu Jusu, Reza Sixo Safai and Michaela Ternasky-Holland.

ADVERTISEMENT

San Francisco start-up OpenAI first showcased Sora in February through a few clips that appear to have the same level of immersive quality as many Hollywood productions. The clips offered a glimpse of woolly mammoths stomping through a snowy meadow, a monster gazing at a melting candle and a Tokyo street scene.

The TFF alumni were educated about the AI tools and were given early access to Sora to create films on their own terms. The filmmakers were required to follow the agreements negotiated with the DGA, WGA and SAG last year with respect to the use of artificial intelligence.

Sora, which takes after the Japanese word for “sky”, builds on the tech behind OpenAI’s image-generating DALL-E tool. Based on a user’s prompt, it expands the information into a more detailed set of instructions and then uses an AI model trained on video and images to create new videos.

It has the potential to become an inexpensive method for creating videos and there is also the fear of creating online disinformation, making it difficult to distinguish the real and the fake at a glance.

Other AI companies too have debuted video generation tools but most of these can produce a few seconds of footage and that too bearing little relation to the prompts. OpenAI, on the other hand, can already create up to 60 seconds of highly detailed scenes.

Google, for example, showcased its new tool, Veo, at its recent I/O developers conference. Filmmaker Donald Glover was present to showcase the tool.

There have been reports about OpenAI CEO Sam Altman and chief operating officer Brad Lightcap warming up to Hollywood executives about what Sora can do.

The overtures come close on the heels of last year’s months-long strikes in Hollywood.

OpenAI has been previewing Sora in a controlled fashion because there is a cost-saving element involved. Besides Google’s Veo, Sora also has to compete with Runway, Pika and Stability AI.

OpenAI recently struck licensing agreements with Vox Media, The Atlantic, People owner DotDash Meredith and NewsCorp. “We’re excited for their short films and eager to learn how we can make Sora a better tool for all creatives,” said OpenAI COO Brad Lightcap.

Follow us on:
ADVERTISEMENT