#information #epistemology #ai # [[Epistemic status]] #shower-thought # Related # TODO > [!TODO] TODO > [[Availability cascade]] [[Information cascade]] # AI generated information is not necessarily evil ![[DALL·E 2022-06-17 21.05.14 - Mephisto standing next to his brother Diablo and Baal, by Wojciech Siudmak, oil on canvas.png]] One of the most dangerous [[Existential risk]] facing humanity is [[Unfriendly AI]], but it won’t be like the movies killer robots. Neither like in [[The Matrix]], less hardware involvement, The Matrix can be created without putting humans in tube. >But if thought corrupts language, language can also corrupt thought >~ [[George Orwell]] [[Language]] is enough to put humans in a mental tube, [[The Matrix is reality]]. [[Artificial intelligence|AI]] generated [[Information|information]] can show a false [[The Map is not the Territory|map of the territory]], but how likely it is? If an [[Artificial intelligence|AI]] wants to control humanity, why not. Why an AI would want to control humanity [[Philosophy/Epistemology/Knowledge|knowledge]] a la [[George Orwell]]'s 1984? Or use humans as energy generator a la [[The Matrix]]? Everything on [[The Internet]] is generated, even without AI, no humans type pixels one by one of an image. On the other hand, [[Artificial intelligence|AI]] can change the semantic space of the [[Information|information]] and produce paradoxical [[Information|information]] (in 2022) such as: ![[Pasted image 20220616175238.png]] How so? Simply by knowing what is a horse, what is moon, what is an astronaut, a human can draw this too. [[Escher]] can do this too, no [[Artificial intelligence|AI]] needed ![[Pasted image 20220616175602.png]] ![[Pasted image 20220616175622.png]] Some people only believe what they can directly experience >Sending people to the moon in the 1960s could have been a Hollywood production. If you limit what you believe to only things you can directly experience, and you are not an astronaut, then a flat-Earth model is what you will end up with. >~ [[Jeff Hawkins]] I think this mindset can become almost reliable in a world generated and potentially hallucinated (false [[Belief|belief]]). ## It's all about [[Epistemic hygiene]] The issue with [[Artificial intelligence|AI]] generated content is that most people have poor [[Epistemic hygiene]], that is, they have a lot of ease in believing what they hear, they are not [[Skepticism|skeptical]] enough and trust what is said on the TV for example. >The ability to observe without evaluating is the highest form of intelligence. >~ [[Krishnamurti]] This [[Artificial intelligence|AI]] generation wave might create two kinds of persons: - The zombie, [[The Matrix is reality]], absorbs [[Information|information]] without [[Do not rely on authority|challenging authority]] - The more [[Skepticism|skeptical]], see that this new pattern of information [[The Map is not the Territory|does not reflect the territory]], thus increasing his general [[Skepticism|skepticism]] # External links - https://mullikine.github.io/posts/the-death-of-personal-truth-due-to-gpt3/