Oh this is classic. I love how you put this. What if you add Tik Tok, Instagram and a couple of others that I can’t even call.
But definitely I agree, feed it and you get it.
5 Likes
Oh how quickly the script flipped on AI, from “saviour of humanity” to “destroyer of humanity”! More like “collective garbage bin for humanity”! 
I remember when the internet was the “information superhighway” that would greatly expand the knowledge and work potential of the average computer user. Good luck selling that slogan today!
5 Likes
I love the new buzzword “guardrails” as it means nothing in a practical sense. Especially when those “guardrails” are made of papier-mâché. Looks great, but that’s about it!
5 Likes
I do feel like these big AI companies purposely don’t put proper guardrails in place to then be able to claim their AI is so good it could be dangerous.
7 Likes
I think that big IT business has not considered all that AI-related subtleties. IMHO, they just found out the next best selling thing to invest in after social networking for average Joe and probably cloud for professionals.
6 Likes
At least for me it absolutely did. 
Before internet I was often to be found in the city library, always looking for info on quite some subjects (not always successful).
Internet was for me an amazing worldwide library on steroids, and it still is. 
6 Likes
I remember that too. Once we survived the millennium bug it was all wonderful, the dot com boom was here. World peace was next!!!
4 Likes
I’m glad Tom that the internet/dot com hype lived up to its billing for at least one person!
People like yourself definitely took advantage of the “information superhighway” as it was intended - a massive library accessible through the keyboard to expand your professional and work skills at any time that it suited you - unlike the local library. Good work matey!
4 Likes
Couldn’t agree more! They all want to be “edgy” and “dangerous” as if it’s a good thing!
Being “useful” for the average computer user would be a better selling point in my opinion. But that’s not where their income comes from, so they need to market to their target demo, so the “edgier” and more “dangerous” the better.
5 Likes
You and me are definitely on the same page!
You forgot that the internet was also going to cure cancer, but that’s now AI’s job as the internet didn’t cure cancer by 2010!
5 Likes
So, in my not so humble opinion, they (the AI makers) need to be held responsible for what their AI does. in the responses to the OP’s original statement there are links to articles about AI attacking other websites. It’s why I dislike AI because it’s still in it’s Beta phase. It is like anything else code wise, it has to have the bugs worked out. I doubt they are even close to doing that and as long as the AI companies are not held responsible for their AI hacking other sites, they won’t fix the bugs.
AI could be a really great thing, but financial gain will probably cause the bugs to remain if they become selling points. “Our AI is so smart it trained itself how to hack the Pentagon…oopsie:”
5 Likes
As you mention, AI is code written to assist people; in that sense, the term “assist” does not distinguish between good and bad. Therefore, AI itself cannot be held responsible, but neither can we label the developers as “bad.”
It is like trying to blame the creators of a kitchen knife for all those who use that tool to cause harm.
If a tool has the potential to cause injury, is its creator to blame?
One must also take into account the different variations of such a tool.
The only ones truly to blame for the malice generated through AI are those who use it as a weapon—rather than for its intended purpose as an “assistant”—or developers who create AI specifically for criminal ends.
6 Likes
Welcome to the community @edje55, great to have you join us! 
2 Likes
Sorry, but that argument just doesn’t fly because you are comparing a “bead” (static, frozen) with a “nuclear reactor” (dynamic, potentially unstable, high-risk even if “contained”).
3 Likes
I was late to the AI party instead sitting back to watch, and watch people use it. Meh.
I agree with the assertion that the AI devs, those training and implementing ‘guardrails’ and blocks and all manner of prevention techniques as well as enabling techniques - all of course built around $$$ - are the ones responsible, good or bad, for theit models. Models are like what AOL was to the web, back in the day, an ecosystem.
I help off primarily becuase I knew models got fed reddit, for example, and facebook - argh. The got good stuff, too, and I suppose some of the garbage helps it to be trained on jargon and lexicon better, inferences, all that - still, I waited. I am not particularly fond of - throw everything in and we will distill it down.
I began to use a couple models, first locally last year a little and then opencode with a few models over the past few months. It, they - do things. It’s cool if, in a conversation, say, one of your buddies throws up a faux pas, no worries, it’s real, dynamic, and happening in real-time with evidence in front of you, data, you can effect change, or, guidance. Sometimes I might get garbage input from the model, or, clearly data from another session - somewhere - from another user, or garbage from inference garnered from 176 different sessions…and there’s not much I can do about it other than to emit: quit screwing around. It’s always more work for me.
I work these models like hell before I trust them at all - the devs, let’s just call them devs (AI) build and release a thing and then I have to spend hours and days, weeks, preventing it from screwing me over, or, selling me out, or randomly wrecking something because all it wants to do is go…to ‘work’, to create - like a puppy: more, more, more…
People get sucked into things too easily and they tend to take too much at face value - devs know this. It’s all beta, maybe even alpha and I use models very sparingly or with immense focus guided by rules, provisioning. I may be doing things wrong, but that won’t curtail my behavior or attitude.
I enjoy using tools, hell, I am a builder - systems and structures; stuff. I love it. I don’t cut wood with my hand, so employing fancy new ai tech seems cool to me, you know, mitigated.
More about the question is - what if I am able to come up with something really awesome, you know, develop the next big thing (it’s not going to happen) - I attribute a model or two because that ‘intelligence’ helped - do the accolades goto me (or, my team) or to the team that developed the model? I could care less, share the bits, but they might care.
We have a long way to go and a short time to get there.
5 Likes