Thanks for sharing this! Noted that it was released the same day the Raine v OpenAI lawsuit was filed.
I point back to my comment about protocol for researchers working with human subjects. Signs of self-harm warrants us to breech confidentiality to have a professional assess the situation; it's actually our responsibility to do so. Thus I fail to see why we can't do the same with someone chatting with AI. If the answer is "it violates the terms and conditions", it seems sensible to change the terms and conditions.
If others have insights that can add some clarity on nuances that I'm missing, please share. I am honestly interested in learning what I may be failing to see.
The spin from openai (the day the Rain v case was filed) seems like a pretty disingenuous attempt to build a counter narrative:
https://openai.com/index/helping-people-when-they-need-it-most/
Thanks for sharing this! Noted that it was released the same day the Raine v OpenAI lawsuit was filed.
I point back to my comment about protocol for researchers working with human subjects. Signs of self-harm warrants us to breech confidentiality to have a professional assess the situation; it's actually our responsibility to do so. Thus I fail to see why we can't do the same with someone chatting with AI. If the answer is "it violates the terms and conditions", it seems sensible to change the terms and conditions.
If others have insights that can add some clarity on nuances that I'm missing, please share. I am honestly interested in learning what I may be failing to see.