The Surprising Websites Included in ChatGPT's Training Data

Expired Data

The use of expired or removed data in AI models raises intriguing questions about data ownership and exclusivity. Jaeden Schafer highlights that 30% of the data used in AI models is no longer available online, making it exclusive to the companies that collected it, like Google 1. This exclusivity could increase the value of such data, especially as platforms like Reddit begin charging for data used in AI training 1.

30% of all the content they collected is now essentially exclusive to them. Right. They have like 30% of it is gone now.

--- Jaeden Schafer

The implications of this are significant, as it means that a substantial portion of the training data for AI models is not accessible to the public, potentially impacting transparency and fairness in AI development 2.

   

Data Concerns

Data exclusivity in AI training raises concerns about ownership and access. Jaeden Schafer notes that much of the data used in AI models, such as personal blogs, is not properly categorized, leading to potential issues with personal data being included without consent 2. This lack of categorization is partly due to concerns about privacy and the inadvertent inclusion of personal information.

A lot of these AI models, it has been said, are actually when they categorize all of the data that they're collecting, they're not really categorizing a lot of it.

--- Jaeden Schafer

The discussion also touches on the ethical implications of using data that is no longer publicly available, as it may lead to monopolistic control over valuable information by a few entities 2.