🌏 In this week’s edition of ‘As the AI World Turns’ we find OpenAI sidestepping the question of what data they used to train their video generator, Sora.
I mean, is there really a question?
My take: Of course they used YouTube videos, and recall YouTube is owned by a major AI competitor – Google.
In an interview with the Wall Street Journal, OpenAI’s Chief Technology Officer Mira Murati, first said she “didn’t know” if YouTube videos were used.
Then she said Shutterstock videos were used.
Then she backed off further and said she just wouldn’t comment but “public” sources were used.
This raises two questions:
1. Would Google ever have the nerve to sue OpenAI if they could prove YouTube videos were used?
Google faces fewer lawsuits over its own training data, but does have one high-profile case it’s dealing with, brought by the Clarkson law firm in California.
2. Why aren’t more lawsuits against Google emerging while OpenAI and Microsoft are fielding several?
Perhaps potential filers feel the case against Google is more difficult given they do have language in their privacy policy stating they use public data to train their models.
Though they did update that language to be more explicit AFTER they were training Bard/Gemini.
What a tangled web is being weaved here by Big Tech…
Comments