Some will argue that ChatGPT was not at fault in a teen suicide, "there were obviously many factors at play". But consider this: if you know you can design for better understanding, and add positive friction that helps people recognize when they are being misled by synthetic text, why wouldn't you?
Find A Helpline | Free emotional support in 130+ countries
For clarity: As was rightly pointed out to me on fedi, ChatGPT is software and is by definition not a responsible entity. At fault are the humans working at and managing OpenAI. While the design of software can harm, the humans behind that design should be more clearly addressed.