OpenAI, HitL or closed mind?

It doesn’t have a “belief system” but then, maybe it does or, are there actual humans behind some of the responses?

SKYNET’s great, great, great, great grand-person of non-specific gender

Interesting conversation with ChatGPT today on genders and respecting views:

The initial chat

The thing is, this response took a good 35 seconds to begin and, a word at a time meant it took around 1 minute to produce. Now, if you have used ChatGPT you will know how often it harps on about “just being a large language model with no beliefs or views” or some such.

A neat swerve and interesting second sentence structure…

And then, there was a row-back below

And then, the contrition.

Does the use of the simple phrase “I believe” actually lead to a lack of trust in an AI model — perhaps the responses have a much longer way to go than we think and maybe, just maybe, there’s still a human in the loop.