NewYou can now hear Fox News articles!
The February 2025 report shows that AI logic models lack a moral compass in the February 2025 report. They will cheat to achieve their goals. The so -called larger language models dels (LLMS) will incorrectly represent the degree of social standards.
None of these should be amazing. Twenty years ago, Nick Bostrom experimented with an idea in which AI was asked to make paper clips most effectively. Given the command and the agency, it will eventually destroy all lives to make paper clips.
Isaac Asimov saw this in his “I, Robot” Stories that consider how the “aligned” robotic brain can go wrong in the way humans harm.
The moral/moral context in which AI runs logic models is kindly small. (Getty Images)
A significant example, the story “runnearland”, puts a robot mining tool on the planet Mercury. If two men on the planet have to return home, they need to work. But the robot order is caught between the demand and the demand for maintaining themselves. As a result, it is circles around inaccessible minerals, not knowing that in the big picture she is ignoring her first order to save human life.
The imminent AI-powered unemployment economy: Who will pay tax?
And the big picture is the point here. The moral/moral context in which AI runs logic models is kindly small. In that regard includes the written rules of the game. It does not contain all the unpopular rules, as if you do not manipulate your opponent. Or that you do not lie to protect your own interests.
That AI may not contain numerous moral considerations in the context of logic models that make humans or AIs from every decision. That is why morals are harder, and the complexity the situation is, the harder it gets. There is no “you” in AI and there is no “me”. There is only prompt, process and response.
So “do with other people …” really doesn’t work.
AI is changing the business. This is how we stay ahead of China
One moral compass in humans develops by socialization, living with other men. It is an incomplete process. Nevertheless, it has allowed us to be in wide, diverse and extremely complex societies without destroying ourselves
The moral compass develops slowly. Humans take years from childhood to adulthood to develop a strong spirit of morality. And many people still rarely get and their fellow is constantly threatening humans. Humans have taken millions to develop our ability for destruction and self-destruction adequate morals. Never work to keep the rules of the game. Ask Moses, or Muhammad, or Jesus, or Buddha, or Confucius and Mensius or Aristotle.
Can a well -aligned AI calculate the effects of its actions on thousands of people and societies in different situations? Can it calculate the complex natural environment on which we all depend on? Right now, even being the Best and being cheating can’t differ. And how can they do? Fair pain cannot be reduced to rules.
AI can’t wait: Why do we need motion to win
Maybe you remember the experiments that the Capuchin monkeys seemed to be “unequal pay” to do the same thing? Was it? When it comes to morality, this develops more than any AI.
It is clearly difficult to see how AIs can be given the sense of socialization of morality and absent of continuous evolution, for which there is no human training for the ability of current models. And even then, they are Trained Town Composition. They are not becoming moral, they are just learning more rules.
This does not make AI useless. It has the enormous ability to do good. But it makes AI dangerous. It demands that moral humans create a guide we create for any dangerous technology. We do not need to run towards AI chaos.
Click here for more Fox News Opinion
I ended up with a bite for this comment, it is entirely based on the events that have been publicly reported. But after a reflection, I realized two things: first, that I was using someone’s mishap for my mic-drop moment; And second, the people involved may be harmed. I left it.
It is immoral to use the pain and suffering of others to further one’s selfishness. It’s something human, at least most of us know. It’s something that AI never understands.
Click here to get the Fox News app
