Chat GPT passed Turing's test. And that means humans cannot separate it from the real human. That thing is one of the biggest steps for the AI. But when we think about ethics and morals, and we try to make sure, that person is a machine or human, we must realize that AI can make answers, that are right from the point of view of morals and ethics.
The idea is that the AI searches lawbooks, to find something moralistic and ethically right answers to the questions. The Western laws based is on morals and ethics, and following the law the AI can tell a person if something is right or wrong. The modern AI has more data where it can find answers than anytime before.
AI can be an impressive tool for code-breaking. It can run the code-breaking program on multiple servers. The system can verify data from different sources.
That means the modern AI can play humans very impressively. The modern AI can make many things faster than humans, but those very complicated systems are not generative. They can connect things from different sources, And that makes it possible for the AI to seem very intelligent. Modern robots can read books, and they can also understand or follow commands that people give. Following orders doesn't mean that the machine understands what it does.
And if the person who controls those systems understands, what something means, that can turn everything dangerous. If we think about robots that operate in some adventure parks and play enemies in some adult fantasy combat tracks, those kinds of systems can turn very dangerous. If somebody gives real guns to those robots' hands, they can be used as robot soldiers. This is one case, where those innocent-looking questions or work can turn deadly.
When we talk about the AI, we must realize that it exists. If we don't tolerate those things and learn to use them right, that causes a situation in which somebody will develop those things anyway. The need to control the AI is understandable. But the control must not deny the development of those things. The problem with defense solutions is that they should be trusted and dangerous. Without those abilities, the AI is unable to complete its mission.
https://www.livescience.com/technology/artificial-intelligence/gpt-4-has-passed-the-turing-test-researchers-claim
Image:
https://dawn.fi/uutiset/2024/06/15/tekoaly-lapaisi-turingin-testin
Comments
Post a Comment