We often hear the term “AI ethics.

Typical problems caused by AI malfunction are as follows

Many problems with AI appliances
AI speakers ordered goods without permission; AI vacuum cleaners wrecked furniture. Smart appliances leaked privacy information.

The company’s decision-making process after deep learning, such as loan screening and stock order placement by AI, was not understood, leading to trouble.
An AI algorithm that determines credit limits for Apple’s credit card (operated by Goldman Sachs) in the U.S.? ゙ Preference was given to men over women. Misplaced stock orders have also disrupted the stock market.

Pedestrian Fatalities Caused by AI-Equipped Uber Self-Driving Cars
An AI-powered car in a suburb of Phoenix, Arizona, U.S.A., was identified as a “human” until the very end. The car bounced without stopping.

AI drones could be used as a mass casualty weapon.
The AI may not be able to identify whether the person with the gun is “trying to hunt or fight,” and may misidentify him or her as either a civilian unrelated to the terrorist act or a person who is “trying to fight” the terrorist act. ゙ Potential for damage.

The biggest problem with AI is that we often do not know “why it made the decision it did.
Perhaps the greatest human task for AI is to come up with “unexpected scenarios”.

We need to come up with “possible” things from “what is not there,” rather than analogizing things from the materials available. In order to develop the ability to envision things for which there is absolutely no data, it is important to be in contact with nature, where human experience and intuition can be developed.

◆ Reference
AI that will destroy mankind!
https://ainow.ai/2019/10/01/176315/

Movie “Westworld” U-NEXT
https://video.unext.jp/play/SID0036898/ED00209438