If the word scaffolding was used more than 31 times per 4 paragraph, it is AI.

Oprindeligt skrevet af Realigo Actual:

If the word scaffolding was used more than 31 times per 4 paragraph, it is AI.

And you should block anyone using the word scaffolding for the first time in a thread.

Done.

Yeah, heard all of these before.

Thing is, what these analyses really highlight the most is just why none of the bots that are truly trying to fool people use ChatGPT.

It's much more difficult to detect a bespoke custom model with single-voice training material - and those will overtake the corporate slopmodels for everything other than the zero-effort content farmers within probably 5 years.

Wow! The topic of spotting AI generated text --- something really hard to do --- is very fascinating! You must have a keen eye for cool facts. Would you like me to write you a short essay about this topic, or is there anything else I can help you with? Perhaps we can spot AI generated content together --- what an adventure that would be![5]

Oprindeligt skrevet af OoOoOoooOOoOoorgle:

Yeah, heard all of these before.

Thing is, what these analyses really highlight the most is just why none of the bots that are truly trying to fool people use ChatGPT.

It's much more difficult to detect a bespoke custom model with single-voice training material - and those will overtake the corporate slopmodels for everything other than the zero-effort content farmers within probably 5 years.

Agreed specialized models would be able to deceive & lie to the average person more, than a generalist corporate slop model will be able to do

Bad thread, OP blocked.

Thank you so much for sharing that video on how to spot AI-generated text — I really appreciate it! It was genuinely helpful and gave me a clearer understanding of the subtle differences between human and AI writing. I wouldn’t have found such a useful resource on my own, so I’m really glad you took the time to send it my way. Thanks again for looking out for me and helping me learn something new!

I always trust YouTube videos when their descriptions are nothing but advertisement for NordVPN.

This comment was sponsored by Raid Shadow Legends.

Thanks for sharing this! It's super helpful, especially with how advanced AI text generation has become. One thing I've noticed, though, is that AI sometimes gets a little repetitive or overly formal. It can also struggle with really nuanced context or humor. If you’re ever unsure, checking for those little quirks can be a good indicator. Looking forward to diving deeper into the video!

— Really great breakdown of key AI red flags!
— Loved how it highlighted both subtle and obvious signs of AI writing.
— The tips on cross-checking for repetition or odd phrasing were spot on.
— The examples were super clear and easy to follow.
— Definitely going to use some of these strategies next time I’m reviewing text!

I am often accused of using AI because my writing style tends to be very similar.

Oprindeligt skrevet af Commander Makara:

I am often accused of using AI because my writing style tends to be very similar.

Same here, but im not really shocked, since the datasets were likely trained off comments like ours.

:shrugs:

Oprindeligt skrevet af EASY PETE:

Thank you so much for sharing that video on how to spot AI-generated text — I really appreciate it! It was genuinely helpful and gave me a clearer understanding of the subtle differences between human and AI writing. I wouldn’t have found such a useful resource on my own, so I’m really glad you took the time to send it my way. Thanks again for looking out for me and helping me learn something new!

......This is a pod person.