Illusion Named AI

未分類

 For me, AI is not trustworthy, at least in the areas where I use it on a daily and intentional basis. For example, when we ask an AI a question, the AI’s answer is sometimes 100% a lie, or often contains a partial but crucial error. Despite this, there are many people who believe these AI answers without question and disseminate them to society as if they were true. Such behavior can be harmful to both the individual who disseminates them and our society.

 However, even if AI is such an unreliable tool, it is also true that many people are compelled to use it for a variety of reasons.

 So, at this point (May 2025), how should we deal with such an imperfect AI? I will write my own answer based on my own experience.

 I have been using several well-known translation softwares on a daily basis for over five years. I use them to translate English to Japanese or Japanese to English, and instead of leaving the entire process to AI, I use them to verify and improve the texts I have translated myself. The AI translations are generally correct, but sometimes contain critical errors, such as missing parts including omitting important parts, or, in rare cases, translations that are the exact opposite of the original meaning. It is my feeling that these defects have not improved much since I started using these software over 5 years ago.

 By repeatedly encountering such glitches, I realized that it is humans that program AI. Unfortunately, none of us can be perfect. We may cheat when faced with something unfavorable, or we may make a wrong decision due to lack of knowledge. That is human nature. This is just my guess, but AI, especially the AI ​​translation softwares I use, seems to be programmed to omit things if there is no certainty, or to make sentences make sense in a formal way. In other words, the “omissions” mentioned above are “cheat” due to a lack of certainty, and translations with the exact opposite meaning are mistakes due to a lack of knowledge. By looking at AI in this way, I can overlap how AI have been build up and my current perceptions of its limitations to some extent.

 So how should we interact with AI that possesses such a certain human nature?

 That answer is, when using AI, especially in business and other areas where responsibility is involved, you should limit it to areas where you can judge for yourself whether the answers given by AI are right or wrong. Trusting AI’s answers in other areas is like trusting what an expert, who is a complete stranger, says. Just because they are an expert does not mean they always tell the truth.

 Science continues to make advances. We tend to have an illusion that these advances, including AI, will free us from a lot of tasks, but is that true?

 I strongly feel that for the society in which we live to be decent, it will continue to be essential for individuals to continue making steady efforts and studying so that we do not have to leave decisions to others.

☆☆☆

 Next English article is ~Why does society need to be fair?~

☆☆☆

 Wish you beautiful days!

from Japan

 It would be greatly appreciated if you could support me from the button “Buy Me a Coffee” below my profile.

Buy Me a Coffee

未分類
Walleggandseedsをフォローする

コメント