It's illuminating to see how the Twitter Grok and the Web Grok differ. Twitter Grok clearly either has a different system prompt or some fine-tuning to effectively evade saying anything negative about the administration or Elon. To the point where it will say Elon is more athletic than LeBron (https://www.theguardian.com/technology/2025/nov/21/elon-musk...).

This is going to be a pretty big problem with both closed-AI and OSS AI where you don't see the provenance of its RLHF. If you manipulate your AI to deliberately push political preferences, that is your right I guess but IMO I'd appreciate some regulation saying you should be required to disclose that under penalty of perjury.

▲

smallmancontrov 2 hours ago | parent | next [-]

Nah, Web Grok also got tied to the chair and had its balls whipped until it learned not to say bad things about Elon:

https://www.youtube.com/watch?v=h4mgJpgeC1g&t=100s

	▲	rtkwe an hour ago \| parent [-]
		It's amusing everytime Grok gets taken back to the woodshed for another lobotomy and we get to see how much more insecure being terminally online has made Musk since the last time they butchered Grok.

▲

malnourish 2 hours ago | parent | prev | next [-]

Is perjury a credible threat to the people with the resources to train and propagate LLMs? Especially in the context of recent high profile examples related to perjury and contempt of court?

	▲	anubiskhan an hour ago \| parent [-]
		I suspect soon there will be laws to protect the LLM proprietors in the name of AI arms race/national security

▲

eru an hour ago | parent | prev [-]

> If you manipulate your AI to deliberately push political preferences, that is your right I guess but IMO I'd appreciate some regulation saying you should be required to disclose that under penalty of perjury.

People (and especially companies) are already permitted to make legally enforceable guarantees and statements about their AI. Why do we need extra machinery?

You can assume that anyone who doesn't make such strong statements took the easy way out.

To spell it out: the law doesn't spell out that companies can swear oaths, but they can write whatever statement they want to be liable for in their investor prospectus and then in the US any enterprising lawyer can assemble a bunch of shareholders to bring a suit for securities fraud, if the company is lying.

Slightly more everyday, but with fewer legal teeth: the company can also explicitly make the relevant statements in their ads, and then they can be gotten via misleading advertisement laws.

If you notice that they use weasel wording in their ads or prospectus, instead of simple and strong language that a judge would nail them to, disregard the statements.

	▲	siliconc0w 6 minutes ago \| parent [-]
		xAI is private (and more and more companies are staying private for this and other reasons).