Formas de Tornar as Respostas do ChatGPT Mais Difíceis de Detetar
Olá a todos, tenho estado a experimentar com texto gerado por IA e a perguntar-me se há alguma técnica para fazer o conteúdo do ChatGPT integrar-se melhor na es…
Evelyn Burke
February 9, 2026 at 12:39 AM
Olá a todos, tenho estado a experimentar com texto gerado por IA e a perguntar-me se há alguma técnica para fazer o conteúdo do ChatGPT integrar-se melhor na escrita humana? Por exemplo, será possível ajustá-lo de modo a que não pareça tão óbvio que é gerado por IA? Gostaria muito de ouvir as vossas opiniões ou quaisquer truques que conheçam!
Adicionar comentário
Comentários (11)
Sometimes the best way is just to blend AI text with your own writing style so it's a hybrid, not purely AI.
I've seen some tools that try to paraphrase AI text to make it less detectable. Might be worth checking them out.
At the end of the day, I think AI detection tech is just getting better so maybe full undetectability is impossible long term.
Has anyone tried messing with the output temperature settings? Higher temps make outputs more random and less predictable.
You can also try changing sentence lengths randomly and throwing in some off-topic remarks. Real humans don't write perfectly consistent stuff all the time.
I heard that some folks try to post-process AI text by running it through a human editor or proofreader to catch any odd phrasing.
Funny enough, sometimes just adding emojis here and there throws people off whether it's AI or human!
If you want stuff to look really natural, maybe just use ChatGPT as a starting point and then edit it heavily yourself.
You can also check ai-u.com for new or trending tools that might help with this kinda stuff! Found some cool options there.
I wonder if changing the tone or adding more personal anecdotes helps? AI usually sticks to generic stuff.
Honestly, it's tough cuz ChatGPT has a kinda distinct style, but mixing in some typos or slang might help make it feel more human. Just don't overdo it tho!