ChatGPT thread

[Jacob]

Yeah the concerns about PIP and proprietary information are nearly universal to larger organizations, so it stands to reason that there'll be attempts to address that with enterprise solutions.

One of the challenges, I expect, is that since what happens inside ChatGPT et. al. is largely opaque to the average observer, it seems like you'll have to simply rely on the guarantees of the vendor that sensitive information of whatever kind isn't compromised. Because how can you know?

On the other hand, I also expect that AI solutions offered by the Chinese will include methods for the surveillance state to detect trends in population sentiment and leverage information for the benefit of the party (including identifying dissidents). I wouldn't be totally surprised either if some level of access could be arranged for major intelligence services from whatever countries the companies providing the solutions originate.