First Sora Short Films and Artists against GenAI Music

Author

Philip Mohr
April 05, 2024

Hey, it’s Philip. TGIF!

Today we are covering OpenAI’s new voice cloning model, video artists playing with Sora and music artists waking up to challenges for the music industry + a lot more!

For new readers, sign up here

THIS WEEKS RUNDOWN

  • Sony adds C2PA to select Sony Alpha cameras via a firmware update. Link

  • EU asks Meta and TikTok to crack down & label AI deepfakes. Link

  • US and UK partner on AI safety. Link

  • Kate Middleton video goes viral based on rumors that its created using AI. Link

  • Deepfake detector by TrueMedia.org launched in Beta

  • YouTube adds tool to label AI-generated content that “looks realistic”. Link

  • Argil.ai deepfakes go viral on social media. Link

TOP STORIES

'Air Head' Demonstrates Sora's Potential in Filmmaking

In "Air Head," the creator group shy kids presents an impressive showcase of how Sora could revolutionise filmmaking. Air Head is just one (more here) showcase of OpenAI's initiative to engage with production studios .

Opting for a "yellow balloon" as the protagonist cleverly navigates around the current constraints of Sora, which can’t reliable recreate objects or scenes from previous prompts.

Nevertheless, the result is striking, demonstrating the possibilities that lie in assembling a short "film" by piecing together video segments generated from various prompts.

Top Artists Unite Against AI's Threat to Music Creativity

Drake, Billie Eilish, and Katy Perry, among over 200 artists, have signed a letter titled “Stop Devaluing Music” through the Artists Rights Alliance, urging AI developers and tech companies to refrain from using AI in ways that infringe on artists' rights and devalue their work.

The letter highlights the potential of AI to undermine creativity and replace human artists, citing instances of AI-generated songs mimicking artists and viral content.

While some artists are exploring AI's boundaries in music, there's a significant call for ethical guidelines to protect artists' livelihoods and intellectual property​​. Source

Platform like suno.ai really showcase the ease of creating compelling songs with AI by just typing a single line prompt.

OpenAI unveils impressive Voice Engine to clone voices

Credit: pinsystem.co.uk

OpenAI has unveiled a new AI-powered voice generation tool called "Voice Engine" that can mimic various voices that are showcased with a few examples. You can check them out here.

Its not just the outputs that are impressive, but also the fact that it only needs 15 seconds of audio to “learn” a voice.

While OpenAI’s model is not out in the wild yet, other platforms like Elevenlabs and respeecher can already be used by anyone.

DEEPFAKE OF THE WEEK

TO CHECK OUT

Reads & Reports:

  • Can AI-Generated Misinformation Be Stopped? - Techopedia

  • India’s struggle with deepfakes ahead election? - TheHindu

  • The Verification Handbook. PDF

Apps:

  • Suno.ai lets anyone create songs from a text prompt with impressive results.

  • TrueMedia’s Deepfake Detector launched in beta and allows anyone to paste a URL to a video and analyse for AI generated contents.

  • Hive Moderation AI-detector for Text & Images.

  • Clone yourself with Argil.ai’s text to video tool to create social media content.

Disclaimer: I used Reword AI to rephrase, spellcheck and improve my writing.