"Open"AI is scanning its users' chats and reporting them to the cops.

futurism.com/openai-scanning-c…

reshared this

in reply to Dan Gillmor

Ah, how this will fail spectacularly! What will actually happen is racial and class-based profiling leading to baseless arrests, and people will still commit ChatGPT-assisted suicides.

There is no amount of RLHF you can put on a synthetic text extruding machine to make it safe. To make calls like this reliably requires understanding of the human condition, and that requires consciousness.

Any impersonation of human emotions & intent by a chatbot should be illegal.

WallOffTrump reshared this.

in reply to Dan Gillmor

We, at PANAfricanSSI, like Perplexity. ai and recently paid for the service. Hopefully, we can receive additional benefits to the general reasoning that it is best at. If it shines at coding when used in VSCodium+Roo Code, that would be awesome. We equally use OpenRouter+Cluade(Desktop).

It is irrational to think that user data would not be trained to improve performance of these models and agents despite that their developers provided an opt-out option.

#engineering