And I'm also aware of the ethical and social implications of artificial intelligence, and the need to ensure that AI models like myself are used in ways that are beneficial to society as a whole.

Sherry Turkle and Joseph Weizenbaum would scream at you for this. There is no "awareness" and the only ethics are external. AI as we have seen it is nothing more than a look-up table of stereotypes and cliches so constraining its behavior, for ethical purposes or otherwise, comes down to careful curation of the training data. Which OpenAI definitely didn't do, and no you can't ask them about that, if you want to see into their process attack them on Twitter.

Cathy O'Niel said that "weapons of math destruction" have three characteristics:

- They are opaque

- They are widespread

- They can do damage

I think it's interesting that Wolfram Alpha has, since its inception, shown you what prompts it derives from your conversational text and offers you syntax to improve your queries. Sidney/Bing, on the other hand, says "you are a bad user 😠". And I think it's sad that doesn't take much hype before everyone acts like the thing in the box is alive.