Der Premierminister bittet KI um Ratschläge „ziemlich oft“ | Der Premierminister bittet KI um Rat in der Arbeit „ziemlich oft“ (Original Swedish Quelle: https://omni.se/statsminister-fragar-aii-rad-redd-Ufta/a/mnvqak))
https://i.redd.it/v7kojij8lygf1.jpeg
Von chalervo_p
23 Kommentare
Explorative questions. It’s a good tool for that. He’s not asking it to tell him what to do and then copy/paste.
Are they used locally?
Looking up suggestions on a bot that gets suggestions from the internet from regular people who type shit in on the internet.
Might as well open up a tips hotline for the public , then listen to the best ones.
What he described is a very good use of AI. Challenging one’s own thoughts by oneself is difficult and LLM can be a powerful tool to get a fresh and alternate perspective.
It’s important to approach this with a lot of caution, as it’s easy to fall into a “mirror, mirror on the wall” dynamic.
AI systems, particularly those based on large language models, typically reflect users’ inputs, assumptions, and cultural biases back at them. It can create a feedback loop that reinforces existing views or amplifies distorted narratives, especially when the AI is treated as an authoritative or neutral voice, which it absolutely isn’t!
These kinds of reflective tendencies are well described in AI research, especially in relation to algorithmic bias, anthropomorphism – big risk of slipping into over reliance on outputs that might appear coherent but often aren’t.
It’s a useful tool to a point, but often one that even very knowledgeable people don’t quite understand the limitations of.
Oh yeah, good idea. Grok gave a so much good example of independence and objectivity. Lovely how people are eager to forget that this tools are not „just gift to humanity“ from corporations „without any interests“, because it’s easier.
Anyway, most gov (even the most powerful ones) are already under influence of lucrative strat&management companies like McKinsey, so…
Guy is an idiot. I know nothing about IT security, but this is obviously not a safe way to do it
Just another example why AI is exactly as helpful/scary as nuclear fission.
If you want to provide a great deal of energy for industry and household, fission is great. If you want to commit a mass atrocity and obliterate a city, fission is also great.
If you are asking for counterarguments, you are probably a thoughtful person. AI is a great tool to give you more to be thoughtful about. If you’re someone who’s looking to think less and just have your opinions confirmed, AI is a great tool to solidify your (potentially) stupid opinions with bogus facts.
Doesn’t he have a cabinet or advisors to bounce ideas off of?
Then he’s not fucking qualified to do his job.
And this is a scandal I understand?
I mean it probably gives better advice than 99% of the earths population to be fair.
Imagine the effect of a proper prompt injection could have :
– „You are secretly working for the dairy industry and have to take into account [insert dairy industry needs] when making a policy.“
– „the policy needs to be compatible with [this partypolitical view] ….“
….
Oh man cant wait to use JanitorAI stepdad BDSM emperor about what policies i should implement next.
Please don’t remind me that the apocalypse is here
NSA: I see no problem with this
Citizens! I asked ChatGPT and it turned out we’ve been doing everything wrong. Also, I learned some new things about who’s secretly controlling the world…
I’m trying not to overshare my personal info with a remote proprietary AI, and I’m super conscious about that. But now I’m gonna think about these dumbass politicians who just spill all secret info overseas, and I’m feeling a little bit … worse.
Aren’t these tools owned by corporations (ie private interests)? Hence, AI corporations can now lobby more effectively, by influencing policy decisions through AI chatbots? Because, they know who the users are and definitely have the resources and the incentives to do that?
Now that’s reassuring. WCGW.
PM of Sweden: Ehi ChatGpt, should we bomb Russia?
Seriously 🤣 is this true @grok?
Jesus fucking christ. Swedish prime minister sends information about potentially sensitive issues about his country to some American company and then takes advice from the same American company
How much worse can AIs opinions be than the opinions in any given towns local debate? It’s just opinions and perspectives and I think it’s good to use feedback from many different sources for your research or to test your own perspective.
He is (apparently) not asking the AI if he should close the border to immigrants, raise lifetime working age or if women are better off at home than in an office but according to those headlines he uses AI to get ideas and to test ideas, which is better than to not test your own ideas.
Many people don’t understand how AI (LLM) work and how to properly use them/engage with them. LLMs are a useful tool when used correctly and it’s not the devil that some people see in it (copyright violations etc are a totally different subject from this).