The Bittersweet Dream of AI That Is Smarter Than Humans

At a glance

This is a column by Kim Dae-sik, a professor at KAIST. He says that building AI with intelligence beyond humans could bring big benefits, such as curing diseases and solving hard problems, but people could also lose the ability to judge and decide for themselves. The writer explains this two-sided nature using the phrase "bittersweet."

Key points

1. 1

   Humans have always built tools to fill their limits, but general artificial intelligence (AI that thinks like a person across many fields) could take over even the judgment of tools.

2. 2

   An AI that grants wishes could produce unintended results, so it is hard to define exactly what we want.

3. 3

   The writer argues we should see AI as a companion rather than a servant, and says he wrote this piece together with a chatbot.

4. 4

   He warns that if superintelligence (AI that surpasses humans) emerges, people could become recipients of conclusions rather than companions who discover things together.

Why it matters

As AI takes on more decisions, we need to think about how much we should hand over to it and what we should keep for ourselves.