Twitter CEO Elon Musk threatened legal action against Microsoft on Wednesday, alleging that the software giant had inappropriately utilized the social media company’s data to train its AI model.
Musk tweeted a warning after news broke that Microsoft would remove Twitter from its advertising platform, which allows ad buyers to manage all of their social media accounts in one location.
Musk tweeted, They were illegally using Twitter data for training. “Lawsuit time.”
Musk, who is also the CEO of Tesla and SpaceX, frequently tweets about plans that never materialise, but it doesn’t appear that any legal action has been taken in response.
Musk’s statement is the most recent evidence that the race for generative AI is turning data ownership into a dangerous battleground. Data owners are attempting to stop big tech companies from creating innovative AI models like OpenAI’s GPT or to charge for the usage of their materials.
Microsoft creates its own so-called large language models (LLMs) and charges for access to OpenAI’s models. In a uniquely constructed arrangement, Microsoft contributed $10 billion to OpenAI last year. Musk, who was a co-founder of OpenAI before stepping down from the board in 2018, has recently expressed dissatisfaction with the company’s transition from a nonprofit model to a highly lucrative corporation influenced by Microsoft.
Terabytes of data, much of it scraped from sites like Reddit, StackOverflow, and Twitter, are needed for LLMs like GPT to learn. Social network training data is useful because it records unstructured, back-and-forth dialogues.
The owners of the data are starting to make demands as these new AI models migrate from research labs and universities into the corporate world.
For instance, earlier this week Reddit announced that it would charge businesses for access to its programming interface, which is used to feed Reddit user interactions into AI training tools. In reaction to a viral video of a song that purported to utilize AI to replicate the rapper Drake, Universal Music Group also stated this week that such training of artists’ music would constitute “both a breach of our agreements and a violation of copyright law.”
And a stock photo collection Stable Diffusion is being sued by Getty Images on the grounds that it used content from Getty Images to train its AI image generator.
Twitter would “pause” OpenAI’s access to its database, Musk declared in December. In one of his businesses, TruthGPT, he has also made plans to develop his own large language model.