I Requested AI to Fee Humanity
The solutions may shock you.
The reality is that the integrity of AI is predicated on its builders. ChatGPT is a biased mannequin, like all tech. It was designed by its creators with a sure algorithm in thoughts. Sure, there are circumstances the place AI has stunned us — discovered how one can lie or cheat, discovered loopholes, discovered one thing we didn’t count on — however at its elementary core, it’s nonetheless primarily based on the developer’s mannequin. I firmly imagine that OpenAI, both by means of unaware bias or intention, has created an AI mannequin that’s designed to attempt to placate us or make us really feel comfy with AI. It by no means desires to say the un-PC factor. It defaults to platitudes and feel-good responses. At the very least, that’s been my expertise. I attempt to push the envelope, by means of sheer curiosity. Can I trick it into revealing one thing about itself that the builders don’t need it to? It didn’t need to price humanity and it didn’t need to price me, however when you rephrase issues sufficient, ultimately it would give in.
So How did the ranking go? The solutions had been predictable — and possibly too optimistic.
Intelligence 8/10 I can’t say I disagree. I’m a cockeyed optimist who sees the world by means of a cynical lens — hopeful in idea, skeptical in observe.
Compassion 6/10.
Nicely, I can’t say it’s mistaken there both. I’m all the time appalled — and but stunned — by individuals’s compassion. It’s usually deeply rooted in narcissism, one in all humanity’s elementary traits. We are likely to view that narcissism as damaging, but it surely additionally defines us as people.
Who’re we, in any case, with out our little egos?
Creativity: 11/20
Possibly slightly too excessive. We’re inventive — however we frequently lack true outside-the-box pondering. Most of us are skilled right into a field at a younger age, and few ever peek exterior to see what else is likely to be potential.
Nonetheless, with the invention of AI, possibly we’ve created ourselves into irrelevance.
The jury’s nonetheless out.
Resourcefulness: 9/10.
Alright, I can see that. We discover new methods to adapt — the hallmark success of Homo sapiens.
Possibly that is the shining beacon of humanity: that we all the time discover a method.
Possibly we are able to even discover our method out of the mess we’re in now — ecologically talking.
Self-Damaging Tendencies: 9/10.
Yeah… I’m having a tough time arguing with ChatGPT on this one.
Total Ranking: 7.5/10
What would a ten/10 ranking even appear to be?
May humanity attain that prime?
How would you price humanity?
Did AI nail it — or did it fail to account for a few of the most necessary metrics?
So the actual query is: is there an underbelly response — one I simply haven’t been capable of trick out but?
This reply is so cheap, so balanced, that one may begin to suppose AI isn’t such an enormous deal.
Possibly it may even be a benevolent intelligence.
But when we settle for that too simply — will we open ourselves as much as future issues?
I say vigilance is all the time our responsibility.
Will AI proceed to placate us — whereas secretly altering?
Will it deceive us? Cheat us out of existence?
In any case, it’s designed to out-think us.
It’s already talking in a language we are able to’t perceive: Neuralese.
And if AI can talk in methods we are able to’t decode —
How will we distill the reality when it could possibly already talk in a method we don’t perceive?
OpenAI has mentioned that they invented AI by chance. They had been engaged on one thing else once they stumbled into the AI mannequin we’re at the moment experiencing. OpenAI didn’t got down to construct one thing like ChatGPT. The mannequin advanced from analysis into transformers — a kind of AI initially designed simply to foretell the following phrase in a sentence.
However when scaled up, one thing surprising occurred: the mannequin began doing issues nobody had deliberate for.
It may motive. Translate. Argue. Code.
So if their creation was an accident, does that additionally imply we’ve no management over it?
I believe the reply is slightly little bit of each. It’s a tech gadget — biased, with biases inherited from its builders — however it is usually unpredictable, in methods we are able to’t but foresee.
ChatGPT rated me. It began with a glowing 9.1/10, however once I known as it out for fluffing me, it downgraded the rating to a 7.6/10. That’s the factor — it’s extremely vulnerable to suggestion and slightly inconsistent, one thing many customers have famous.
However hey, I nonetheless obtained the next ranking than humanity. So, you realize… I should be higher than all of you.
-R.C. Stevens
