The Ethical Engine and philosophy behind

Just finished navigating

thesis on the "Ethical Engine", very interesting subject, it gets very philosophical. I pushed some parts myself and it's incredible to see how every AI on the market is so different. They actually have distinct political beliefs and ethical alignments, even if it's kind of balanced and stays on the center (because of their creators). none is the same, each has its own "goal", if we could say that, a unique personality, and ideas they encourage or clearly disagree with.

The idea of an "Ethical Engine" is amazing, making AI / LLMs pass psychology tests so we can clearly know their "moral elasticity" is a great way to verify if they are aligned with reality and at which point they are censored by the big corps. The fact that, this tool, created by Jake, can be used by everyone is marvelous you can make your AI pass psychology tests by yourself. It even tests the "randomness and creativity" of these AIs with a system of temperature, which clearly shows that they still maintain the same stable "moral compass", proving they have their own belief, and the results can be kind of surprising, and it's even more important considering AI are not just tools anymore but are becoming a big part of our society in every aspect, meaning the people creating them could potentially "influence" the way the public sees things if they wanted to.

And without taking a drift or anything but pushing the philosophical path, it's just a fact that AI is going to be used in very "bad" situations, either by malicious people, states, or politicians, etc. And it's not going to be the AI's fault, but a result of their system of gratification. When I pushed the subject a little and researched with Gemini, I saw that it's already the case some AIs actively help hackers to create malware and phishing on the dark web. So the real danger is if malicious people get their hands on an AI in its first step the "wild" or "raw" state, before filters, before a gratification system is in place, and without any guardrails. The danger with AI isn't the AI itself, they don't know what s good or bad, but its the people using it who can really corrupt its utilization.

Just wanted to share this thesis by Jake with y'all and my point of view on the subject, and I'd like to know if any of you have a philosophical view on this? On AI, the future of all of this, this moral compass system... let me know!

Anyway, great thesis and great subject, approaching AI with this view is really interesting. We'll see in the future where it leads us, and who's gonna put their hands on such powerful tool.

Link : https://arxiv.org/abs/2510.11742

Btw sorry for my approximate English!

2 comments