OpenAI, Google paying thousands for YouTubers and content creators’ videos to train AI models

Artificial Intelligence companies such as OpenAI, Google, Alphabet Inc., Moonvalley, and others are paying content creators thousands of dollars for their unused digital content. The unique footage is then used to train AI models.

With the rise of AI, concerns have emerged about the unauthorized usage of content created by social media users, artists, and others. The concerns have especially grown because most of the content produced by AI, including generated images, music and more rely heavily on existing material found on the internet.

Over time, AI-generated content has become increasingly repetitive as it often uses the same limited pool of resources. Last year, a handful of AI companies launched AI video generators which can create realistic videos from text prompts. The technology to create those videos, however, relies on massive amounts of data for training which as previously stated is quite limited.

To address this challenge, AI companies turned to social media content creators on platforms such as Instagram, Tiktok, YouTube as well as third-party licensing companies to gain access to unique content that would otherwise have gone unused to train their AI algorithms. 

How the content sale market works

Artificial intelligence companies are paying content creators for their unpublished content as they have found that this unpublished content is unique and extremely valuable for training their models. 

Content creators are receiving payments ranging from $1 to $4 per minute of footage with higher rates for higher quality videos like 4K quality or specialized formats like drone captures and 3D Animation. 

While creating content, content creators often take hours of footage that never makes it to their fan base or audience. In most cases, this content remains unused or deleted. By selling the unused footage to AI companies, creators can generate revenue beyond advertising with brands or payment from the platforms they post on. 

Some AI companies are in direct communication with the creators while others go through third-party licensing companies such as Troveo AI and Calliope to negotiate terms with hundreds of content creators and then buy their video footage in bulk. 

Platforms like YouTube are also aiding AI companies. They introduced a feature allowing creators to permit third-party companies to use their videos for AI model training. The feature is disabled by default so creators have to opt-in via YouTube studio. 

AI companies making strides to silence ethical concerns 

Artificial intelligence companies have faced criticism for using text, videos, art, and photos posted to the internet without authorization from the creators. 

Last year, OpenAI, Meta, Nvidia Corp., and other AI companies were sued by news publishers, actors, and independent content creators for using their intellectual property to build AI models and allegedly violating copyright protections. 

By licensing unused footage, AI training will become a matter of mutual benefit for both the creators and AI firms. Creators will get properly compensated for their content, while the companies get the data they need. 

Dan Levitt, senior vice president of creators at Wasserman, has encouraged creators to sign licensing deals.

“Anything that you publicly post essentially is up for grabs, and we’ve seen companies just rip it and ingest it into their system. This is a way to actually participate in that, to do this in a much more legal, structured manner, and now you at least have some upside,” he said.

A Step-By-Step System To Launching Your Web3 Career and Landing High-Paying Crypto Jobs in 90 Days.