Can Artificial Intelligence rule or defeat humans by itself?

TLDR: No! But authoritarians use AI to build controllable destructive or policing machines.

The “intelligence” in Artificial Intelligence is misleading. Intelligence is inferring information, converting it to knowledge, and applying the knowledge to adaptive behaviors¹. Computer algorithms, with human supervision, can collect data and form knowledge (e.g., statistical learning). However, humans feed data into algorithms and instruct them on how to construct and apply knowledge. Also, machines are not alive and have many deficiencies. For example, machines cannot reproduce, and you can unplug them! So, it’s not today’s concern.

But, imagine we are in the year 2122, and technology has kept the same pace for the last 150 years. Scientists made computers that can reproduce by building other computers. They have auto-heal systems and are cooperative. If you unplug a machine, it will fight with you and uses its battery as energy to plug itself back. Even if you dismantle a device, other intelligent machines will fix it so you cannot easily destroy them.

In this scenario, someone should build such computers, and it’s not a one-person job. It needs groups of scientists, engineers, and leaders having a specific goal with sufficient funds working years on multiple research and production lines. Thus…