All we can make are assumptions, because we don't see how throwing our intellect into a pile helps against life severance, what gold in which hills you're talking about, or what range of intellectualism and brains is going to ram us into which unspecified dangers or how to advise about any of this, because there's is absolutely nothing being said in this string of words. There is no concrete statement in here at all. You never make a point and tell us 'no' when we grapple with it.
Your flow of conjecture is a glassy-still body of water reaching unblemished to the horizon, painted in colors of unease and anxiety, and you're asking us which fallen trees to watch out for. There is no tree nor wind nor iceberg nor landmass to turn eyes on and you're getting frustrated that we're imagining shapes on the horizon.
You want to talk about AI, talk specifically about AI. Current implementations of AI, current threats of AI, reasonable speculation based on current implementations. You want to talk about the limits of human intelligence, you find the research about it and make a point about it, because at this point I can only guess that maybe there's too many books out there or perhaps there's threats posed to our mental health based on the inundation of advertisement we receive daily (there's plenty about this) or if we're just going to buffer overflow on data and our brains will explode (they don't) or perhaps we're just going to run out of music, having played every reasonable permutation (sidenote, we won't).
What's a modern problem? What's 'modern?' My linoleum floor is modern relative to hay thatching, my linoleum floor problem is that there's a squishy spot by the kitchen.
If you need an intellectual aphorism to cool your jets, here's this- one may be too smart for their own good, and curiosity may have killed the cat, but satisfaction brought it back. The more successes you make, the more mistakes, sure, but you try, you fail, you try again, you fail better. Sucking at something is the first step to being kind of sort of okay at something. Or, to take the power out of failure and blunt the language of the previous statements- mistakes and errors are not failures until you refuse to correct them, and best not to let that stop you from reaching for your goals- better to ask for forgiveness than permission, yeah?
RE: EuchreJack, there IS a good point to be made here- computers 'lie' to us all the time in order to give us a better user experience. Progress bars and loading bars are pretty egregious examples, since they usually don't exactly reflect the progress on an operation or accurately represent how much progress is made, but we get happy brain chemicals if the CPU is waiting to read the disk and the progress bar is still moving on a load or copy operation. Your phone's cellular connectivity strength probably isn't the full five bars it tells you it has all the time because the end user does not need to sweat about their 67% connectivity strength when it won't interrupt service. If you're playing video games online, the action as rendered from your computer are not happening serverside as you're seeing it happen on your screen, the server is making best-guesses about where things should be and where you ought to be respective of your movements and actions and interpolates where things should be accordingly. Hiding digital dumbness from people makes things go smoother. The scope of digital deception does not extend past decisions made by developers who rightfully understand end-users are idiots and don't want them complaining about things that don't actually matter. This isn't man vs. machine, this is developer vs. idiot.
Now, if I wanted some concern about technology being leveraged against people, I'd be reasonably enough concerned as a US citizen about the NSA's PRISM program and its other surveillance programs on its own nation and people. I'm not an important person, so besides knowing it's shitty, I don't lose sleep over it. I'm generally unhappy with the kind of information aggregation megacorporations like Amazon, Facebook, and Google perform, since I'm important enough (I have a wallet) that they'd want to sell me stuff, but their angle is strictly to sell me stuff, and I can take measures (adblock, tracker blockers, script stoppers, DNS blockers, just not using their services) so as not to be annoyed by the worst threat they have to offer- targeted ads. This is within my grasp to affect, and so I do. The worst of AI we're seeing now are weird swirl paintings based on words fed into an algorithm to aggregate a million pictures with some iota of relation to the prompt, or a markov chain designed to spit out a 'story' (a loose aggregation of words and actions that go together), or, like, CleverBot, or that other CleverBot-esque personality that Microsoft let users play around with and fed it enough input to parrot racist diatribes.
AI as you fear it is ridiculously computationally expensive because it's pulling correlations from millions of sources to churn out singular results and is designed specifically within its scope to produce just that type of result. CleverBot will never write a melody, AI Dungeon will never understand how to build IKEA furniture, wombo art will never so much and send a ping command in a terminal. Facial recognition is thwarted by facemasks and tape, thermal tracking is foiled by glass windows, Tesla autopilot is useless in any circumstance that isn't totally mundane freeway driving and even that can't be trusted to allow drivers to put their focus elsewhere.
But this is the part where you say 'no' and 'you don't get it' because, as before, I'm making examples of things I can only speculate you're looking to discuss, because there's been no point offered to grapple onto, or because I'm telling you not to worry about certain things and/or do not agree with your sense of distress.