Will we find out before 2025 about a nation state using Generative LLMs for generating propaganda messages?
Basic
14
αΉ€2.9k
Dec 31
93%
chance

Still counts if the use or the discovery happened before this market opened. To be clear, this is about using LLMs to generate new text, with applications such as ChatGPT or LLaMA. Other uses of LLMs don't count. This is a longer term version of the market below, and barring anything controversial that happens to that market, I intend to resolve this YES if the market below resolves YES. use of DALL-E-3 or similar image generators that use generative LLMs as part of their pipeline does count, but there needs to be specific evidence that these were used, not other image generators.

Get αΉ€600 play money
Sort by:
bought αΉ€2,500 YES

@Shump this can resolve YES. Since FUH has been banned not sure if you want to not defer to him any more. https://openai.com/index/disrupting-deceptive-uses-of-AI-by-covert-influence-operations/

There's more than one network that should count, but Spamoflauge has been directly attributed to the Chinese government and OpenAI says they used their services to generate pro-China content then posted to social media.

Just wanted to clarify that I will resolve this in line with FUHs judgement, even though I took a big YES position on the other market.

Let's say there's a state known for its high-tech sector, volunteering spirit of its citizens, and the collosal ineptitude of its governmental ministeries. If private citizens do it for the state, does it count?

@YotamFederman hypothetically speaking

@YotamFederman also, what if the propaganda is meant for foreign people? Still counts?

@YotamFederman Has to be made for the state. It's fine if contractors do it, but they should be contracted by the government, not individual parties/politicians. Propaganda for foreign people counts

Wow that cover photo is surprisingly good. Love the robot with the microphone. Curious why it chose to put the British flag there though.

More related questions