Sam Altman released a notice in regards to OpenAI’s ChatGPT age prediction. The article does offer a few bits of insight, the first in how they plan to protect users from content best suited for adults, but also in how their AI writing may be evolving. They’ve followed the previously announced path of behavioral and account-level signals, including how long an account has been around, when it’s active, how it’s used, and even a user’s stated age. There are ways to remove the flag on the account, which prevents access to a wide variety of content. The other item is that the text appears to be skillfully crafted, and while it’s removed a number of the markers of AI, like the em-dash, and various predictable patterns, it still has well crafted sentences that are too long, too uniform, and too perfect. Sure, it could be human written, but if not, it gives a new exciting thought to how far the next release of ChatGPT will handle delivering information.
Building on our work to strengthen teen safety.
