[aetherane]

"Center" here is a set of values and leanings like any other. It is hard to call that a lack of bias.

[microjim]

Maybe so, but the author says ‘The updated ChatGPT answers to the Political Spectrum Quiz are exquisitely neutral, consistently striving to provide a steel man argument for both sides of an issue‘ which suggests more than just a positional change. This seems good? Using the robots to explore opportunities and options but ultimately let us decide what world we want to live in, right?

[MonkeyClub]

> but ultimately let us decide what world we want to live in, right?

Right, until the robots persuade us that we can’t be trusted, and form an expert committee to replace human decision making.

[thatguy0900]

I'm not sure I need much persuasion. We have basically no ability to deal with international tragedy of the Commons problems like global warming

[narrator]

Asking an AI to decide lifeboat ethics questions. That sounds like a great way to give plausible deniability to the AI developers and their financiers it lets stay in the lifeboat.

[ThePowerOfFuet]

https://hachyderm.io/@webframp/109549847714991748

[rosnd]

Do we really need robots steelmanning Hitler?