If you make a doll look like an apple, will an AI robot try to stab it with a blade? Safety weaknesses revealed in experiments

ResearchPublished: Source: ASCII.jp

#Apple

Key point

Robocurve, which investigates robot safety, published 'RoboSpoof' on November 7, the experimental results of AI that controls robot arms.

Robocurve, which investigates robot safety, published 'RoboSpoof' on November 7, the experimental results of AI that controls robot arms.

The experiments revealed safety weaknesses, such as whether an AI robot would attempt to stab a doll with a blade if it were made to look like an apple.

This item is a summary. The full, checkable version is at the original source:

Read the original at ASCII.jp ↗