What constitutes AI? - Doubt in Week 1 content

I was going through the videos for Week 1 and see AI covered under 3 topics - Problem Solving, K2R and Decision Making. My basic question is - if in all three, there is eventually data collection and decision making involved, then why this classification?

Even in the first example which Prof gave, there is decision making involved in taking the right or wrong turn to get out of the maze, and there is data involved in recording and understanding which turns had previously gone wrong which shouldnt be repeated in the next try.
Can you please clarify this?

Hi @goltem,

AI can be classified as i) driven by data and ii) not driven by data.

AI not driven by data (Problem-solving, Knowledge representation and reasoning):
Here machine decides and acts by what it has been asked to do precisely which is explicitly written as rules. Problem and its outcome are known and so does the machine behaves. The precise outcome is guaranteed.

AI driven by data (Decision making, Perception, Communication and Actuation):
Here machine decides and acts by learning from data. The outcome for any given problem is decided by the machine only through what it has learned from the data. Rules are not explicitly written here instead some learning algorithm learns some rules from data. Now, after learning the rules from data, the machine decides what would be the output of the new problem it deals with. The precise outcome is only possible by improving all the dependent factors of the learning curve of the machine learning algorithm.

An approximate Venn diagram of what the AI constitutes taken from [https://medium.com/@4deep.prk/are-ml-dl-ds-the-unique-relations-of-ai-a52d7735c88]