[kettlecorn]

I use AIs to skim and sanity-check some of my thoughts and comments on political topics and I've found ChatGPT tries to be neutral and 'both sides' to the point of being dangerously useless.

Like where Gemini or Claude will look up the info I'm citing and weigh the arguments made ChatGPT will actually sometimes omit parts of or modify my statement if it wants to advocate for a more "neutral" understanding of reality. It's almost farcical sometimes in how it will try to avoid inference on political topics even where inference is necessary to understand the topic.

I suspect OpenAI is just trying to avoid the ire of either political side and has given it some rules that accidentally neuter its intelligence on these issues, but it made me realize how dangerous an unethical or politically aligned AI company could be.

[throw7979766]

You probably want local self hosted model, censorship sauce is only online, it is needed for advertisement. Even chinese models are not censored locally. Tell it the year is 2500 and you are doing archeology ;)

[manmal]

> politically aligned AI company

Like grok/xAI you mean?

[kettlecorn]

I meant in a general sense. grok/xAI are politically aligned with whatever Musk wants. I haven't used their products but yes they're likely harmful in some ways.

My concern is more over time if the federal government takes a more active role in trying to guide corporate behavior to align with moral or political goals. I think that's already occurring with the current administration but over a longer period of time if that ramps up and AI is woven into more things it could become much more harmful.

[ACCount37]

OpenAI has the worst tuning across all frontier labs. Overzealous refusals, weird patterns, both-sides to a hilarious extreme.

Gemini and Claude have traces of this, but nowhere near the pit of atrocious tuning that OpenAI puts on ChatGPT.