Chatgpt answers are unreliable

[Vikjohn]

Few days ago, I asked this question, is sugarcane a fruit?, then I got this answer "Sugarcane is considered a type of fruit. Although it is often thought of as a type of grass or a vegetable, sugarcane is actually a type of giant grass that belongs to the family Poaceae and is classified as a fruit because it contains seeds encased in a sweet, juicy flesh. The seeds, or "cane buds," are surrounded by the sweet, juicy stalk, which is used to produce sugar and other sweeteners. The stalk is also used for animal feed and to make paper, textiles, and building materials."

First it says "sugarcane is a type of fruit, although it's often though of as a type of grass or a vegetable"
Sugarcane isn't a type of fruit at all, and not a vegetable, the rest of the answer is correct but that first part is concerning, it shows ai answers can wrong be sometimes/manytimes, and isn't reliable.

And it can't even solve simple maths, it can also often give incorrect answer if the questions involves dates (history).

 

[Thunder Digitals]

Sugarcane isn't a type of fruit at all, and not a vegetable,

Then what it is?

 

[Vikjohn]

Then what it is?

Like I said, the rest of the answer is correct, it's actually a grass, and not a fruit or a vegetable at all, I'm just trying to say the answer can be unreliable, so using chatgpt for school work or professional work is risky.

 

[iBuySMM]

There are a lot of issues with that gpt that people ignore. This is the least of my concern with it.

 

[Thunder Digitals]

Like I said, the rest of the answer is correct, it's actually a grass, and not a fruit or a vegetable at all, I'm just trying to say the answer can be unreliable, so using chatgpt for school work or professional work is risky.

I was joking around.

Yes, you shouldn't totally rely on chatgpt for those kinda works. You should use it to get some help or an idea.

 

[SOCIAL MEDIA MARKETING]

According to a study published on the bioRxiv website in December 2022, the results were that 32% of AI citations were mistakenly written by humans and 14% of human citations were mistakenly written by AI. The AI quotes are quite vague and create a "recipe" feeling.

 

[Elvlin]

same with answer from google you need to double check it. even when 99% of it's citations is proven to be correct.
seem critical thinking will becoome rarer in the future

 

[brianfulan]

same with answer from google you need to double check it. even when 99% of it's citations is proven to be correct.
seem critical thinking will becoome rarer in the future

Yes, I agree that the convenience provided by AI in the future will make people lazy to read and think critically about the answers they receive.

And more and more hoaxes will make the knowledge we receive biased.

 

[Vikjohn]

Yes, I agree that the convenience provided by AI in the future will make people lazy to read and think critically about the answers they receive.

And more and more hoaxes will make the knowledge we receive biased.

Yeah, even know people have become lazy, I remember back then in school, when calculator isn't a thing, we were made to memorize the multiplication table, now adays, kids can't even do simple maths, they do everything with calculator.

 

[SOCIAL MEDIA MARKETING]

Yeah, even know people have become lazy, I remember back then in school, when calculator isn't a thing, we were made to memorize the multiplication table, now adays, kids can't even do simple maths, they do everything with calculator.

Yes, Technology always has 2 sides, but don't overuse them or even depend on it, we should master technology.

 

[Iwearglasses]

Like I said, the rest of the answer is correct, it's actually a grass, and not a fruit or a vegetable at all, I'm just trying to say the answer can be unreliable, so using chatgpt for school work or professional work is risky.

Glad you informed us on the sugarcane. Saved me from a lot of embarrassment .

 

[DeadJoe]

It will never be 100% reliable, it is a bot, not a human. It is just there for convenience.

 

[GenesisOne]

It is just a tool - not the oracle - lower your expectations mate

 

[Vikjohn]

It is just a tool - not the oracle - lower your expectations mate

I'm wasn't expecting anything more, it's a great tool, just trying to create an awareness

 

[Konyagi]

ChatGPT is still a baby. It has still got a long way to go in replacing human thinkers, researchers, and writers. I pray it doesn't eventually do so because...where will our kids get the motivation to research and discover new ideas if some overrated robot is left to do all the work?

 

[HenryObi]

Few days ago, I asked this question, is sugarcane a fruit?, then I got this answer "Sugarcane is considered a type of fruit. Although it is often thought of as a type of grass or a vegetable, sugarcane is actually a type of giant grass that belongs to the family Poaceae and is classified as a fruit because it contains seeds encased in a sweet, juicy flesh. The seeds, or "cane buds," are surrounded by the sweet, juicy stalk, which is used to produce sugar and other sweeteners. The stalk is also used for animal feed and to make paper, textiles, and building materials."

First it says "sugarcane is a type of fruit, although it's often though of as a type of grass or a vegetable"
Sugarcane isn't a type of fruit at all, and not a vegetable, the rest of the answer is correct but that first part is concerning, it shows ai answers can wrong be sometimes/manytimes, and isn't reliable.

And it can't even solve simple maths, it can also often give incorrect answer if the questions involves dates (history).

Yes, ChatGPT answers can be very unreliable. It's same with other tools as well.

If it's topic you don't know much about, you'd think it's spitting quality and accurate information but when you do a little bit of research at least to verify important information you start to notice that it makes things up.

 

[Pythagore]

Of course, ChatGPT has a lot of small problems.. Anyways, that's just an AI.
It helps people for a lot of different things, tho.

 

[ExperimentalSEO]

I don't like that it is programmed to not say offensive things, and they patch any of the work arounds. It should be neutral and whatever nice or messed up thing it says is because of the instructions given to it.