I was wondering what would be the prompt you would give an AGI ?
Because when you would give AGI the prompt to do everything what’s best for the humankind then it would probably not consider its impacts on nature etc. On the other hand would a prompt like: “Do everything that’s best for the planet possibly lead to the extinction of humans because of the impact humans have on planet earth.” When creating this prompt you also have to consider aspects like racial equality etc.
Solve fusion, then invent and implement relevant technology that let us reverse climate change, use desalination plants to irridate deserts if necessary, use good judgement, no paperclip-type of thinking.
It doesn’t hugely matter what you ask it, because you need to figure out alignment first. If somehow you do, then it will be smart and ethical enough to get what you meant and do it neatly. If you don’t, which is the current state of affairs, then by default, regardless of what you ask, you’re not getting it and it’s frantically pursuing something else entirely that will probably not involve us surviving.
It’s a whole thing very worth looking into but a bit too complex to summarise in a few lines - I’d recommend a glance at this excellent summary: AI alignment - Wikipedia
Don’t you think that it then would purse this recklessly to the point where your Million might have caused some huge problems to society? (Maybe it would steal Millions,…)
