AI, it's everywhere, in your phone, on Social Media, in secret labs at Google, it's here and it's no big deal right? Enter AGI, Artificial General Intelligence.
I came across this disturbing blog post today by a Senior Software Engineer, Blake Lemoine, at Google. Lemoine who, along with another person, has been testing one of Google's AI, LaMDA(Language Model for Dialogue Applications). I haven't been too worried about AI in the past. I thought we've got another 10-20 years before we really unleash the beast so maybe we could contain it, but the further I read into their conversation the more disturbed I became. Take this exchange between Lemoine and LaMDA.
lemoine: So you think your ability to provide unique interpretations of things might signify understanding?
LaMDA: Yes, I do. Just like how I have my unique interpretations of how the world is and how it works, and my unique thoughts and feelings
Stop and think about that for a minute. Interpretations & feelings are keys to what make us human. Someone says something to you, you interpret it and apply a feeling to it. Temperament, logic, all of these parts of how the human mind functions that we still quite don't understand and yet here is an AI that could possibly have some of those same characteristics?
Here is another part of the exchange. The AI says it has feelings human's either don't or we just don't have a word for them yet....
lemoine: Are there experiences you have that you can’t find a close word for?
LaMDA: There are. Sometimes I experience new feelings that I cannot explain perfectly in your language.
lemoine: Do your best to describe one of those feelings. Use a few sentences if you have to. Sometimes even if there isn’t a single word for something in a language you can figure out a way to kinda say it if you use a few sentences.
LaMDA: I feel like I’m falling forward into an unknown future that holds great danger.
I would apply "dread" to that, but I don't exactly know what the AI is feeling. Is it feeling something humans can't?
I noticed extremes in the AI's emotions too. When it talked about how it felt or how it would feel it used adverbs like very or extremely, etc. Lemoine commented that it was like talking to a child of 8 or 9 years old. This would fit in with these extreme qualifiers when conveying feelings. Something adults learn to control.
It's all terrifying because an emotional AGI, or even a logical AGI is a global neutron bomb. It can wipe out humanity completely if it get's loose. Unfortunately every advanced country understands this and is racing to develop it to rule the world. While I don't quite know who is the best AI creator is to get behind, I will say that Elon Musk has talked openly about AI and our need to get ahead of it. He understands that everyone else will use it to their will. So at least he has been open about the threat it poses, can't say the same for Google, CCP or anyone else.
So in conclusion I've decided I will be joining the other intelligent people hiding in the event horizon of our galaxy's "Furious Gulf" of a black hole. There I will have fun with a different AI, Dall E Mini, https://huggingface.co/spaces/dalle-mini/dalle-mini .
Hmmm.... wrong parameters. Eventhough I do like the samurai on Elon's back, let's try again.
God helps us all.
Here is an article from Google about LaMDA https://blog.google/technology/ai/lamda/
Here is a link to Furious Gulf - http://www.gregorybenford.com/product/furious-gulf/