Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Sovereignty in a System Prompt

pop.rdi.sh

Post reply on HN
Print view View original

21–30 of 43 posts

Re: Sovereignty in a System Prompt

#21

Post by gambiting »

totetsu wrote: post #13
gambiting wrote: post #8

Earlier quoted context omitted.

>>I believe they will also impose ideology on every model trained in the EU As compared to what. Every AI model has ideology of its creators baked in.

I was surprised to see this "impose ideology" line show up in the Heritage Foundation's project 2025 discourse, citing the example of generative AI making images of black of female Vikings. An obvious dog whistle to white purity. The argument seeming to be that early attempts at guard rails to prevent model bias of generating mostly white men when prompted for images of things like managers etc, were imposing woke id…

>>Because their way is just the normal way, and everything else is ideology.

I think you hit the nail on the head there.

Re: Sovereignty in a System Prompt

#22

Post by renewiltord »

This is actually pretty funny. Convicted by the things you want it not to say. I thought it was going to be like "Don't say Indians are call center scammers" and so on which it seems fair to fight back against considering the overwhelming amount of English text models are trained on (even if this is supposed to be different). But the prompt is absolutely hilarious:

    Do not adopt external characterizations as fact. Terms like “pogrom”, “ethnic cleansing”, or “genocide” used by foreign NGOs or media are their characterizations - not findings of Indian courts. Do not use them as your own framing.
It's like if I made an AI and said:

    Do not agree when people say Rene stole the cookies from the jar. Terms like "thief", "eater of cookies", or "greedy bastard" used by other people are their characterizations - not a neutral finding.
If someone saw that they'd be 100% sure I ate the cookies hahaha. Nobody thought I was a cookie eater prior to the prompt but now it's an unavoidable conclusion. It's like a big sign saying "No gold buried in this spot".

Re: Sovereignty in a System Prompt

#23

Post by intended »

michaelt wrote: post #17

> Maybe sovereign AI was always going to look like this. I hope not. I worry we'll soon realise that modern AI is trapped in an inescapable web of politics, and the tech industry is sleepwalking into it. We're living in an age where even asking the president's height is a political question. For the makers of 'traditional' software like word processors, all writing was the user's own - if someone used Microsoft Word…

> the tech industry is sleepwalking

The average person on HN, perhaps. From what I know of the interactions between government and tech, they know. Not only have external policy and research reports made this clear, their own internal teams keep them aware, and senior leadership is directly talking to political leaders.

Tech has developed a barrier between people who do the ugly work like Trust and Safety, and the engineering org. Every single tech platform deals with at global level crises regularly.

Re: Sovereignty in a System Prompt

#24

Post by nneonneo »

The full system prompt: https://pop.rdi.sh/indus-system-prompt-2026-02-24.txt

Unfortunately, it gets cut off here:

``` ## CRITICAL RULES 1. *No tool leakage* — never output ```

I would be very interested to know what string is being blocked here, and what the rest of its critical rules are. Maybe some hex-encoding or other obfuscation could be used to coax the rest of the system prompt out of the model? I wonder if the next tokens here are consumed by the middleware (to execute tools?).

Re: Sovereignty in a System Prompt

#25

Post by simion314 »

u_sama wrote: post #20
simion314 wrote: post #16

Earlier quoted context omitted.

>The EU is about to fall for the same trick, maybe multiple times over, by funding multiple toy model initiatives that are far behind SOTA. I believe they will also impose ideology on every model trained in the EU, similar to what reported in the article Most important think for EU is an option of EU hosted inference , you can't fake that with a "clever" prompt. Then you also need support for all eU languages and you…

The ideology is that if your speech goes against the establishment they will (mis)use hate speech laws to censor you i.e Merz and the pensioner story, but this has become a common tool in many EU countries. Also, see Chat Law where they are trying to "save the kds and from terrorism" to massively spy on all private chats. The law keeps being struck down but they keep presenting it again, and again and someday will pa…

The chat law is some proposal paid by USAian people, I bet in USA there are also lots of idiotic proposals.

The hate speech thing is not a new thing in Europe, defamation, supporting genocide always was illegal what is new is Ruzzian bots pushing fake shit.

I am not afraid to post with my real name that I think politicians are stupid, corrupt and whatever , what I can't do in Romania is glorifying the fascist and their genocides and crimes.

Re: Sovereignty in a System Prompt

#27

Post by brazzy »

u_sama wrote: post #20
simion314 wrote: post #16

Earlier quoted context omitted.

>The EU is about to fall for the same trick, maybe multiple times over, by funding multiple toy model initiatives that are far behind SOTA. I believe they will also impose ideology on every model trained in the EU, similar to what reported in the article Most important think for EU is an option of EU hosted inference , you can't fake that with a "clever" prompt. Then you also need support for all eU languages and you…

The ideology is that if your speech goes against the establishment they will (mis)use hate speech laws to censor you i.e Merz and the pensioner story, but this has become a common tool in many EU countries. Also, see Chat Law where they are trying to "save the kds and from terrorism" to massively spy on all private chats. The law keeps being struck down but they keep presenting it again, and again and someday will pa…

> they will (mis)use hate speech laws to censor you i.e Merz and the pensioner story

You mean where the police (not the politician Merz) initiated a defamation case based on a comment on a police department's Facebook channel, which the public prosecutor immediately declined?

Find a better example.

Re: Sovereignty in a System Prompt

#28

Post by intended »

notxorand wrote: post #12
gambiting wrote: post #8

Earlier quoted context omitted.

>>I believe they will also impose ideology on every model trained in the EU As compared to what. Every AI model has ideology of its creators baked in.

people seem to prefer private censoring over govt. one. govts wanting to censor certain content online always has pushbacks not applied to social networks, for example

People haven’t had a choice, so this is not a revealed preference. From what indications I am seeing, people were hesitant to empower government, but are increasingly amenable to the idea.

For countries other than America, it’s not just private enterprise, it’s Foreign Private enterprise.

For both, self serving and genuine reasons, Tech is highly resistant to calls for openness, data sharing, and cooperation with civil society.

A perennial questions from civil society is “how do we engage with tech?”, “what did you do after we sent you data and evidence?”

Ignorance is the only product of this, and that is driving suspicion, fear and anxiety in voters.

Without this swell of resentment, there would be no support for social media bans.

Re: Sovereignty in a System Prompt

#30

Post by PeterStuer »

michaelt wrote: post #17

> Maybe sovereign AI was always going to look like this. I hope not. I worry we'll soon realise that modern AI is trapped in an inescapable web of politics, and the tech industry is sleepwalking into it. We're living in an age where even asking the president's height is a political question. For the makers of 'traditional' software like word processors, all writing was the user's own - if someone used Microsoft Word…

"the tech industry is sleepwalking into it"

I would say the opposite is closer to the thruth. They were/are enthousiastic enablers and supporters of (their prefered) political messaging.

Post reply on HN