3 ms·
Then it is time to take control by seeing such databases with false information. We could probably design data specifically to attract an AI system looking for
by sandworm101 3y ago
Then it is time to take control by seeing such databases with false information. We could probably design data specifically to attract an AI system looking for well-documented data. I wonder how images of Steven Colbert uploaded to imgur it would take to convince the AI that he actually was.
- nwoli 3y agoI don’t think this is a possible approach. It would basically mean needing to feed in so much noise that a human wouldn’t be able discern reality from fiction given no priors. Modern ML is too smart
- BiteCode_dev 3y agoAKA encryption.
- proser 3y agoGiven how easy it is to keep an average human from discerning reality (see the last decade of boomers/gen xers on Facebook as an example) and the massive potential to create slight variations with LLMs, I don't think a Stephenson-style misinformation propagation campaign is that far outside the bounds of possibility. That's a lot of compute power to waste on it, but I would guess that that's what bot networks are going to be used for in the future (or already are, right now, if they're done mining bitcoin).
- sandworm101 3y agoThat's why I said president of canada. A non-intelligent AI would not find any priors indicating who was the president. There isn't one. So a few thousand images in the database, against zero information to the contrary, might be enough.