Are there any guarantees required for AI licensing agreements?

[seonumbershop]

Ahead of its IPO, Reddit CEO Steve Huffman revealed that the company has raised more than $200 million through licensing deals.

Some see licensing deals as a win-win situation: publishers get paid for their data while AI companies get access to vast amounts of quality data.

However, this also comes with some drawbacks. Social media platforms like Reddit are community forums where people can post just about anything. Conspiracy theories, misinformation, hate speech.

The quality of content moderation on Reddit.
While Reddit has moderators and content policies, they didn't ban hate speech china number screening until 15 years after the site was founded. Is this the kind of content AI models should be training on?

AI companies can clean their data to filter out this type of content, but there is no clear standard on which to build each model. So, as a consumer, I would not know what data the models were trained on and how they were "cleaned."

Should we reject certain sites for training AI models?
This then raises the question: Should certain websites be banned when it comes to training AI models? And what safeguards are in place to ensure their models don’t regurgitate the darkest content on the internet?
Voice actors have been the backbone of ElevenLabs' text-to-speech generator since its launch. Now the company is rewarding them by investing in them again.