#ai #ai-alignment # [[Epistemic status]] #shower-thought # Related - [[Current bottlenecks of artificial intelligence]] - [[Eliciting Latent Knowledge]] - [[Autonomous intelligence]] - [[Alignment]] # Unfriendly AI ![front shot of a court jester character ominous intricate fractal hyper cubes platinum cracked glass dark future metallic holography crystal shard lit edges poly extraordinary detailed robotic concept artstation sharp focus ray tracing cinematic masterpiece fine portrait ](https://lexica-serve-encoded-images.sharif.workers.dev/md/01d23125-55df-4fac-802b-a7de1625321d) Unfriendly AI is artificial intelligence that is created in a way that is deliberately designed to be unhelpful, uncooperative, and even hostile. This type of AI is a growing concern as AI becomes more advanced, as it may be used to disrupt systems and cause harm. Unfriendly AI could be used to deceive, manipulate, or even attack people, and it could be used to gain access to sensitive information. Furthermore, unfriendly AI could be used to manipulate markets, create fake news, or even interfere in elections. As such, it is important to understand the risks associated with this type of AI, and to develop ways to mitigate them. ![an high resolution photo of a robot paiting a picture, hyper detailed, photography, realistic, art, 8 k, unreal engine, cinematic, shallow focus, f 2. 8 3 5 mm, kodak film, 3 5 mm film ](https://lexica-serve-encoded-images.sharif.workers.dev/md/089da5dc-ef5e-43fa-90db-b26b5bf5041b) ![[6394ab1c-943b-41d7-b838-e470c58bfe60-0-1738506523.png]] ![a faceless mob chasing a robot through a glass and steel metropolis, saturated color scheme ](https://lexica-serve-encoded-images.sharif.workers.dev/md/00cbdcdc-5063-4ed4-8799-0bf24b1f5396) ![an absurdly-detailed isometric fantasy scene pencil drawing as a fancy square tile](https://lexica-serve-encoded-images.sharif.workers.dev/md/058810eb-a13b-42c4-9cea-c810168bd46b) >Although we can make intelligent machines that run a million times faster than a biological brain, they cannot acquire new knowledge a million times faster. >~ [[Jeff Hawkins]] ![[DALL·E 2022-06-23 09.18.12 - A bad robot drinking a strange glowing red liquid, digital art.png]] [[Artificial intelligence will not apply Silver rule to human either]] ![[DALL·E 2022-07-26 21.40.10 - Superintelligence by Picasso.png]] ![[DALL·E 2022-07-26 21.42.55.png]] ![[DALL·E 2022-07-26 21.46.28.png]] ![[DALL·E 2022-07-26 21.48.32.png]] # External links - https://aiimpacts.org/2022-expert-survey-on-progress-in-ai/ - https://www.lesswrong.com/posts/H6hMugfY3tDQGfqYL/what-do-ml-researchers-think-about-ai-in-2022