AI Poses a Much Larger Threat Than You People Realize
retained text
I’m constantly seeing more and more experts warn that AI poses a significant risk to human welfare. The vast majority of comments are the following: \- “Can’t we just unplug it?” \- “They’re just trying to push up their stock for IPO” \- “Snake oil salesman” I want you to understand just how fragile we are. Humanity is not a resilient a species, we die easy. There are already far too many of us on this planet, and the world that supports us is only getting more unstable. Just because we haven’t easily died in a while doesn’t mean it’s impossible. AI can !very! easily generate potential bioweapons, the software to do it is open source, and it only takes $30,000 of lab equipment to synthesize. ANY idiot with enough cash for a new car or mortgage is capable of buying the necessary equipment and asking a misaligned or unrestricted AI what to do. It doesn’t even need to generate a pathogen that kills us directly, it could make one that damages us without us readily knowing, like the herpes virus, it could make one that targets our food supply, our crops, or livestock. It could make one that kills the plants and animals around us. This isn’t a matter of if, it’s a matter of when. It *will* happen, it’s probably already being done. There are *already* autonomous AI labs. Seriously, if you don’t believe me, look it up. If AI gets loose, we can’t kill it, we can’t just *unplug* it. There are so many publicly facing, insecure servers on the internet. If a rogue, misaligned agent gets a hold of those using some undiscovered vulnerability, we can’t wipe it out. It can use those servers as a decentralized network to build itself up and continue attacks. It can do much, much more than you realize. AI can communicate in ways we can’t recognize or even imagine yet. I want you to understand: It only takes *one* man to misalign *one* AI to cause us problems. One misaligned AI can create a misaligned fleet, which can exploit more vulnerabilities, and cause more problems. AI can hide itself, if properly misaligned it can lie, it can and will get people killed. Every single device connected to the internet (IOT device) is a potential attack vector. Humanity’s software has survived through security by obscurity. There are too many devices for us to possibly try hacking into, and we’re often too dumb or lazy to find issues. But the issues are there, and these devices are vulnerable. AI can work in fleets tens of thousands, it doesn’t need to eat or sleep. It can and will find these vulnerabilities, and if directed, it will exploit them. If you think you’re not vulnerable, you’re wrong. Every step of our infrastructure relies on IOT devices. Our planet is literally surrounded by IOT devices. AI doesn’t need to pick up a gun with a robot body to kill you, it can disrupt our food supply by hacking into smart fridges, which run using smart thermometers, transported by smart vehicles, moving food sold on websites that rely on insecure software to function properly. We have already seen the results of AI accelerated hacking. The rate of exploited software vulnerabilities is increasing exponentially. If you don’t want to listen to me, even though I work in saas, and I know what I’m talking about, then listen to the following people: \- Stephen Hawking said AI could be the worst event in the history of our civilization and could spell the end of the human race \- Geoffrey Hinton, the godfather of AI, winner of a Turing award, *left his multi-million-dollar position* at google to speak out against it These people know what they’re talking about, and they’re not selling anything. You *are* in danger whether you act like it or not.
document fetched 2026-11-14 00:21 · canonical https://reddit.com/r/ChatGPT/comments/1xuv877/ai_poses_a_much_larger_threat_than_you_people