As exciting as an AI future is, some researchers are aware of the risks involved in AI succeeding too well. In 2016, Google’s AI subsidiary DeepMind taught a computer program to play a variety of Atari video games at a superhuman level, in only a few hours. The program was given no background knowledge, which meant it managed to learn the game from scratch.
Questions to discuss in the classroom:
Should these results impress or scare us?
What are student’s ideas on AI and Deep Learning?
Are humans in danger through giving such power to technology?
How will this shape the world as we know it?