Elon Musk's AI startup, X, has updated its privacy policy to allow the use of user-generated content to train its artificial intelligence models. The new policy, set to take effect on September 29, states that X may utilize publicly available information for the purpose of training its machine learning and AI models. This change is not reflected in the company's previous terms, which are still accessible online.
Musk clarified on X that this data usage will only apply to publicly available information and will not include direct messages or other private content. However, it remains unclear how exactly the information from X will be used and which AI models it pertains to. Insider reached out to X for further comment but has not yet received a response.
Interestingly, Musk has been critical of other platforms using X to train AI, even threatening to sue Microsoft in April for "training illegally using Twitter data." This move by Musk to allow the use of public posts for training his own AI models seems somewhat contradictory.
Musk's new AI startup, xAI, aims to use artificial intelligence to gain a deeper understanding of the universe. This launch comes despite Musk's previous warnings about AI potentially leading to the destruction of civilization. He has also expressed concerns about AI systems more powerful than OpenAI's GPT-4, even signing a letter calling for a pause on their development.
Notably, Musk was a key early supporter of OpenAI but left the company's board in 2018. He has been critical of the chatbot ChatGPT, claiming it is "trained to be politically correct" and therefore untruthful.
The demand for large amounts of data to train generative AI products like ChatGPT has led companies to scrape publicly available web sources such as Reddit and Wikipedia. Some platforms, like Reddit, have even started charging AI companies for data access, while others, like The New York Times and Amazon, have blocked OpenAI's web crawler.
Meta, another company developing generative AI models, has recently introduced an option for Facebook users to opt out of sharing their data for AI training. However, there are limitations to this opt-out feature, and complete removal of user information from AI training databases is not guaranteed.
In conclusion, Elon Musk's X has updated its privacy policy to allow the use of user-generated content for training AI models, but it remains to be seen how this data will be utilized and which AI models it will be applied to.