Claim: 

AI's are capable of understanding.

Argument Style: 

Deductive. This style of argument uses a series of building claims that lead to a conclusion, and the only way to dispute the conclusion is to find a problem with one or more of its supporting claims. In short, if you can't find a flaw in the logic, you have to accept the conclusion.

Argument: 

1. When humans debate whether or not AIs have understanding, we should frame the discussion around practical impact to society, not academic nuance.
2. The question that most affects this is AI's potential ability replace, compete with, and otherwise disrupt human work.
3. I define Functional Understanding as the ability for a system to evaluate a problem, find patterns between the problem and the concepts in its knowledge base, and produce a useful output based on that comparison.
4. Both humans and AIs have Functional Understanding.
5. Much of the objection to AIs being able to "truly understand" comes from people—either consciously or unconsciously—using a different definition for understanding, which I call Experiential Understanding. This type of understanding is the sensation of "getting" something, and it requires self-awareness and therefore consciousness.
6. AIs (as of 2024) do not have Experiential Understanding because they are not conscious.
7. Returning to the point of practical impact on society, AIs do not need Experiential Understanding to be useful in society, and therefore they don't need it to disrupt human work.
8. Therefore, AIs are capable of understanding in the way that matters to us as humans.