Is Technology the Best Thing?

  • 107
  • 4
  • 1
  • English 
Feb 23, 2019 10:30
Recently, I had an opportunity to experience machine translation, to be precise, some of the latest automated translation tools. My company is apparently considering to introduce it for improving operational efficiency, and those who routinely engage in translation had to evaluate it to see how it actually works. Consequently, I figured out some interesting points.

I have to conclude that they’re pretty accurate in many situations. In most cases, they’re enough to convey what you want to tell to a large extent. Meanwhile, they don’t really work in some situations. In other words, while they’re often very competent, they’re sometimes stupid. I realized this relates to what I saw in a TV documentary regarding artificial intelligence.

Researchers in the show was trying to create an AI robot that could pass the University of Tokyo, the most prestigious university in Japan. (I was not sure how it could take the exam for humans) They admitted it failed in the end, and the main reason was because it couldn’t understand simple situations we experience in daily life even though it could calculate almost everything from a tremendous amount of data. This is actually what I saw in the process of testing machine translation. And, they conclude that the big difference between humans and AIs(at least current models) is that the AIs don’t understand “meanings,” but we do. If by any chance the day when they could solve this problem comes in the future, I’m anticipating but worried that they would become “Skynet” or dominant machines from Matrix.

I’m going to continue this topic in the next entry.

Thank you for reading.