We Asked GPT Some Tech Questions, Can You Tell Which Answers Are Human and Which Are Not?

[17thDimension]

Let's see if you can spot the difference between a human writer and an AI. We asked GPT 4.0 various tech questions and put them side-to-side with our answers taken from TechSpot explainers.

Read the full article here.

https://www.techspot.com/article/2654-gpt-vs-human-tech-qa/

 

[Thatsdisgusting]

The main difference is that gpt answers are nerdy and tedious, while human answers are more like "Hold my beer I'll explain that to you fast and simple"

 

[Blqckout]

I got 1 wrong, because I didnt even know sata to m.2 existed, thanks, ill buy some

 

[UncleMikeRetro]

I had 10 right out of the 11. Sometimes, it was only a matter of how the explanation started. Humans tend to use different filler words than the AI (so far).

 

[WhiteLeaff]

Easy. AI is very linear and repetitive in text structure. In addition, in most of the answers, it makes final considerations as an opposing point.

 

[VariableSpike]

The structure in which Chatgpt replies is quite similar from question to question, and it tends to use quite formal language and prose (for example, wouldn't expect it to use the word gibberish), one or two threw me off just simply because of the way the question was written

 

[NicktheWVAHick]

I must be a *****. I got all of the answers wrong.

 

[Julio Franco]

The structure in which Chatgpt replies is quite similar from question to question, and it tends to use quite formal language and prose (for example, wouldn't expect it to use the word gibberish), one or two threw me off just simply because of the way the question was written

What if I told you both answers are GPT.

 

[Blqckout]

What if I told you both answers are GPT.

No! T-this cant be!

 

[QuaZulu]

What if I told you both answers are GPT.

What if Julio Franco is GPT?

 

[Blqckout]

What if Julio Franco is GPT?

what if we are all GPT

 

[wiyosaya]

10 of 11. In at least one case, AI did not consider what technology will do. The one I got wrong was, in part, because I have no personal knowledge of the subject.

 

[VariableSpike]

What if I told you both answers are GPT.

Colour me surprised then if that is the case!

 

[Hodor]

I must be a *****. I got all of the answers wrong.

Are you actually ChatGPT?

 

[plplplpl]

11 out of 11, but I've had a professor's career of data training reading student papers.

 

[wiyosaya]

I got 1 wrong, because I didnt even know sata to m.2 existed, thanks, ill buy some

M.2 to USB exists, too.

 

[wiyosaya]

I must be a *****. I got all of the answers wrong.

Interesting.

 

[QuaZulu]

what if we are all GPT

Then the Matrix is real (I just haven't mastered bullet time yet)

 

[godrilla]

What if I told you both answers are GPT.

Artificial intelligence is just the summation of human intelligence. This exercise proves that their will always be a human anomaly and an ai one.

 

[godrilla]

The main difference is that gpt answers are nerdy and tedious, while human answers are more like "Hold my beer I'll explain that to you fast and simple"

Give it time eventually it can adapt to outsmart some humans for eg.

OpenAI's brand new GPT-4 AI managed to ask a human on TaskRabbit to complete a CAPTCHA code via text message — and it actually worked. In other words, an AI just fooled a human into checking an "I'm not a robot" checkbox — a terrifying example of an algorithm being able to escape human control.Mar 15, 2023
https://www.fudzilla.com/news/ai/56542-ai-pretends-to-be-blind-to-fool-captcha

 

[Julio Franco]

Colour me surprised then if that is the case!

No! T-this cant be!

Haha, just wanted to stir the pot a little bit. We didn't lie in the intro of the article, there is a human answer adapted from our own explainers and the GPT answer. Cool exercise nonetheless.

 

[loki1944]

Multiple choice roulette

 

[loki1944]

10 of 11. In at least one case, AI did not consider what technology will do. The one I got wrong was, in part, because I have no personal knowledge of the subject.

AI doesn't consider anything; it isn't intelligent or self aware; just human programming limited by the program.

 

[Benchsalyst]

9/11 , both wrongs on Sata MVe ( Q6 ; Q7 ) , maybe because my bad english , I choose the simplest text answer . Easy to know which one is a computer here because of fancy of human answers .

 

[Benchsalyst]

One of possible consequence of chat GPT or IA that could be included in internet is that participative community can disappear because people doesnt want to have matter to a computer rather than a human