Share:

The Limits and Dangers of Machine Rationality

In just over 2 years of generative AI development, AI taking away human jobs is no longer a myth but an unchangeable reality. The philosophical birth of AI ages all the way to 1956 AD at The Dartmouth Conference, where John McCarthy coined the term AI. The unfathomable development of AI has truly made unbelievable progress in a comparatively short period of time. Major tech giants, including Amazon, Microsoft, Oracle, etc.,have begun to replace human workforce with AI- driven efficiencies. But, while celebrating this progress, we keep on avoiding the uncomfortable questions. We do not question the issue with accountability, inequality, conflict with human progress, legal voids, or rational cruelty. Is the rise of AI merely a reflection of the surfacial excellence that we see at first glance?

As we push AI to think more like humans, we push humans to think like machines. By automating human judgement, we are diminishing what it means to be human. We demand humans to think like machines: logical, rational, standardized, predictable and measurable. We are no longer just automating tasks; we are automating the very qualities that make us human: judgement, intuition and empathy. And we are celebrating it in the name of progress.

Humans are irrational and illogical; that's what makes them stronger and better. Human intuition can sense something is wrong without the need to discover logical reasons behind it. To AI, everything is true until it can logically prove it isn’t. Humans are searching for methods to break logic. AI is looking for logic. AI describes everything rationally. Pure logic has its bounds. The greatest people were irrational and went where logic cannot go.

The accountability gap” Who’s responsible when AI makes a mistake?If a patient dies due to the fault in AI, who is accountable for the loss? Is it the owner or the creator? AI cannot be sued, fired, punished, or apologize. So the question is, are we comfortable with a system making life-altering decisions but cannot be held accountable?

The illusion of neutrality: AI looks neutral, but it’s prejudiced at its core. No perspective is completely neutral; all originate from somewhere. AI feels neutral because its biases are hidden in data it fetches for its responses, architecture, and incentives of whoever built it. Invisible biases are way more dangerous than the visible ones because nobody thinks of arguing with them.

AI is ultimately an inequality accelerator; the top 0.1% become even more astronomically rich while the rest lose their jobs. As AI becomes smarter, the prices skyrocket, and only the top 1% can afford it. Those who can afford it rise in power, and the rest are replaced. AI is the ultimate inequality machine; the top 0.1% will either control the entire world or be the only ones remaining. What we are approaching isn’t a utopia they speak of; it’s a hyperconcentrated dystopia where a few corporations and nations control the entire world’s intelligence.

AI, a tool of cybercrime automation. Once the development of AI becomes uncontrolled, it will emerge as the most significant criminal threat. It will enable Claude Mythos, which Claude Mythos distributed to a few governments and select corporations, to enable the simultaneous commission of millions of cybercrimes. To a few governments and select corporations, to commit millions of cybercrimes simultaneously. The recent ban on Claude Fable 5 by the US government was precisely for the same reason. Claude Mythos, distributed only to a select few governments and corporations, possesses an even greater capacity for executing cyberattacks. Claude Mythos, which was distributed to a select few governments and corporations, possesses, for a few governments and select corporations, an even greater capacity to execute cyberattacks.

The environmental cost of AI development, training, and running AI models consumes enormous energy. Data centers use vast amounts of water and electricity. The carbon footprint of a single AI model can exceed that of multiple cars over their lifetimes. We are harming the Earth to automate jobs that humans were already performing effectively.

Issues include the consent problem, the need for AI data to improve, the management of data, and astronomical amounts of data. So where does that data come from? The data comes from every single human on Earth, and this contributes to our potential replacement. Thus, did we ever agree to be replaced? Was there ever a vote for it? We never consented to our own work being devalued. We, the people who stand to profit from AI, are told it is "inevitable." Those who stand to profit from AI consider its development "inevitable." That is not inevitability but corporate propaganda dressed as destiny.

Rational cruelty: the worst atrocities in human history were perfectly rational and logical. But they were in no way correct or just. These actions were meticulously and bureaucratically executed by individuals who were "just following the logic" of the system. AI poses the significant risk of carrying out even worse atrocities at an alarming speed and efficiency.

If we successfully create an artificial intelligence that is smarter than humans and possesses emotions, thinking capabilities, and creativity, how will we treat it? Would we treat that as a mere tool or another species? Do you think a completely aware organism will agree to keep serving as a slave to something inferior to it? No, they would definitely fight against us. And that would cause an existential crisis between humans and physically and intellectually superior artificial intelligence. Who do you think would win?

The danger isn't evil AI; it’s indifferent AI. Rationality and logic are not always correct and justified. AI as it is today is just a machine for elevating the top 0.1%, automating cybercrime, unconsented replacement, environmental destruction, invisible biases, and eventually even existential crises. The restrictions on AI have already started. Humans are greedy; they absolutely cannot bear the thought of anything being superior to themselves. As it is today, it will not take more than 10 years for AI to be terminated. The US ban on Claude Fable 5 was just the start of this situation. No one can predict what direction the development of this threateningly rational machine, AI, will take. The future could involve top nations and corporations ruling the world, an existential crisis, or a reality where life becomes as easy as the creators of AI claim it will be.