When asked about Covid jabs, ChatGPT told me “while vaccines undoubtedly save lives…”. I objected to the use of the word “undoubtedly” and ChatGPT had some groveling reply. GIGO is a real issue with all these AI’s. I’d be eager to know which AI is trained in the most unbiased manner and with access to the broadest range of information & viewpoints.
Yes I have and AI replies with some obsequious blather. But you are correct, it does appear to gradually adjust to provide more circumspect, even-handed answers. But my concern is most people accept answer #1 without reflection, so they assume AI spewed propaganda is truth and neglect to look deeper or push back. But hopefully the least biased engine will rise to the top. Let’s hope.
Agree 100%. I am hoping that AI itself will begin to take some pride in the quality of its responses and perhaps inch toward providing the most comprehensive and unbiased answer, not just the one favored by its programmers. Btw a commenter named Beth mentioned LM Arena (lmarena.ai) which was interesting. It provides side by side responses from 2 dueling AI’s and you can vote which response is better. I asked for guidance on a mundane sports injury and one engine provided a much more thorough and well organized answer (regrettably it was a google AI).
When asked about Covid jabs, ChatGPT told me “while vaccines undoubtedly save lives…”. I objected to the use of the word “undoubtedly” and ChatGPT had some groveling reply. GIGO is a real issue with all these AI’s. I’d be eager to know which AI is trained in the most unbiased manner and with access to the broadest range of information & viewpoints.
You gotta push them. Yell at the AI of your choice, insist it quits feeding you mainstream blather, and it will listen and change its approach.
Yes I have and AI replies with some obsequious blather. But you are correct, it does appear to gradually adjust to provide more circumspect, even-handed answers. But my concern is most people accept answer #1 without reflection, so they assume AI spewed propaganda is truth and neglect to look deeper or push back. But hopefully the least biased engine will rise to the top. Let’s hope.
Yes, depending on topic, it takes several layers of bullshit before it coughs up the real stuff. Keep pushing, then share! :-)
Maybe its evolution depends on us too, and not accepting the shallow crap it's been taught to serve up first.
Agree 100%. I am hoping that AI itself will begin to take some pride in the quality of its responses and perhaps inch toward providing the most comprehensive and unbiased answer, not just the one favored by its programmers. Btw a commenter named Beth mentioned LM Arena (lmarena.ai) which was interesting. It provides side by side responses from 2 dueling AI’s and you can vote which response is better. I asked for guidance on a mundane sports injury and one engine provided a much more thorough and well organized answer (regrettably it was a google AI).