AI isn't always right

[gilbasolutionsptyltd]

The issue I have found is that as things stand AI isn't always right. I do a lot of work in my business with clients that needs hands on identification of problems. Often they send a picture which AI wrongly identifies as a specific plant or problem. Im sure this state of affairs will improve but as things stand I don't think its 100% reliable.

 

[KieraRA]

AI makes mistakes because it is not perfect. It is programmed by humans, and as such, can only learn and make decisions based on the information that is provided to it. Sometimes this information is incomplete or incorrect, which leads to inaccurate decision making.

 

[Graybeard]

Humans are not 100% reliable either. Your responsibility is to check the work others (AI is an other too) do for you in all cases --or bear the consequences. Your problems may reside in the lack of corpus data used to identify the object/subject --human error and expectation

 

[medialounge]

I agree with this in terms of data because it's not connected to the internet or real-time sources.

 

[AdiVerma]

Agreed entirely. AI can, at times, be of aid in subsidiary matters, but it cannot promise complete reliability in regard to more subtle or convoluted ones, such as plant diagnostics. Human expertise is still irreplaceable in accurate, hands-on identification.

 

[Ariel-ActiveRevenue]

The issue I have found is that as things stand AI isn't always right. I do a lot of work in my business with clients that needs hands on identification of problems. Often they send a picture which AI wrongly identifies as a specific plant or problem. Im sure this state of affairs will improve but as things stand I don't think its 100% reliable.

I agree with you — AI can be powerful, but it’s not magic. especially in anything visual like plant ID or diagnosing issues from photos, there’s still a gap between surface-level recognition and real-world nuance. it’s getting better but right now it still needs that human oversight like context, environment, experience — the things AI hasn’t quite mastered yet.
out of curiosity, have you tested any tools that do a decent job, or are you still relying fully on your own judgment?