GPTZero rates the text as 100% human, for what it's worth.
March 28, 2025 at 8:50 AM
GPTZero rates the text as 100% human, for what it's worth.
GPTzero does better but still, I mostly see screenshots from the above
February 16, 2026 at 10:56 AM
GPTzero does better but still, I mostly see screenshots from the above
Some fun things are happening in this thread. One user points out Jimbo's message may be AI generated according to a commonly used AI detector. Jimbo denies the claim.
November 3, 2025 at 4:42 AM
Some fun things are happening in this thread. One user points out Jimbo's message may be AI generated according to a commonly used AI detector. Jimbo denies the claim.
As just one example: GPTZero, one of the most popular detection services, is owned by Superhuman AI (text generation), who also owns Humanize AI (detection avoidance).
August 29, 2026 at 9:32 PM
As just one example: GPTZero, one of the most popular detection services, is owned by Superhuman AI (text generation), who also owns Humanize AI (detection avoidance).
has anyone just run the thing through gptzero
May 29, 2025 at 11:56 PM
has anyone just run the thing through gptzero
Un moment va peut être falloir que les profs soient formés et se renseignent sur le fonctionnement de la détection de gptzero hein.
February 13, 2025 at 10:43 PM
Un moment va peut être falloir que les profs soient formés et se renseignent sur le fonctionnement de la détection de gptzero hein.
GPTZero says the blurb on the website is 100% AI written as well. It also says Truss's statement is written by a human.
December 8, 2025 at 9:35 AM
GPTZero says the blurb on the website is 100% AI written as well. It also says Truss's statement is written by a human.
perplexité de GPTzero va être basse : le texte lui semble familier, donc probablement généré par IA.
A l'inverse, si la probabilité est basse, alors la perplexité est haute, et GPTzero est "surpris" de voir ce mot : c'est sans doute un humain qui l'a foutu là.
A l'inverse, si la probabilité est basse, alors la perplexité est haute, et GPTzero est "surpris" de voir ce mot : c'est sans doute un humain qui l'a foutu là.
February 13, 2025 at 11:53 PM
perplexité de GPTzero va être basse : le texte lui semble familier, donc probablement généré par IA.
A l'inverse, si la probabilité est basse, alors la perplexité est haute, et GPTzero est "surpris" de voir ce mot : c'est sans doute un humain qui l'a foutu là.
A l'inverse, si la probabilité est basse, alors la perplexité est haute, et GPTzero est "surpris" de voir ce mot : c'est sans doute un humain qui l'a foutu là.
Heureusement que GPTZERO précise cela sur leur site...
December 6, 2025 at 11:10 AM
Heureusement que GPTZERO précise cela sur leur site...
GPTZero believes this is 100% AI generated, which makes it even more pathetic.
They’re using AI to generate insurrection fan fic.
They’re using AI to generate insurrection fan fic.
January 26, 2026 at 6:52 PM
GPTZero believes this is 100% AI generated, which makes it even more pathetic.
They’re using AI to generate insurrection fan fic.
They’re using AI to generate insurrection fan fic.
Superhuman, which also has an AI detection tool as part of Grammarly, has snapped up GPTZero.
Superhuman acquires AI detection startup GPTZero | TechCrunch
Superhuman, which also has an AI detection tool as part of Grammarly, has snapped up GPTZero.
techcrunch.com
June 23, 2026 at 9:51 PM
Superhuman, which also has an AI detection tool as part of Grammarly, has snapped up GPTZero.
The results come from the company GPTZero, which claims they “uncovered hundreds of AI-hallucinated citations that slipped past the three or more reviewers assigned to each submission, spanning at least 53 papers in total.”
NeurIPS papers contained 100+ AI-hallucinated citations, new report claims | Fortune
An analysis of NeurIPS 2025 papers by startup GPTZero reveals how AI-generated citations are slipping into elite academic research.
fortune.com
January 21, 2026 at 4:27 PM
The results come from the company GPTZero, which claims they “uncovered hundreds of AI-hallucinated citations that slipped past the three or more reviewers assigned to each submission, spanning at least 53 papers in total.”
o @anarcofino.bsky.social falou na live de hj que a declaração da defesa do Tércio Arnaud era puro chatgpt
joguei no gptzero e kkkkk
(sim, sei que tem falsos positivos, mas nesse caso fodasse né)
joguei no gptzero e kkkkk
(sim, sei que tem falsos positivos, mas nesse caso fodasse né)
November 21, 2024 at 8:31 PM
o @anarcofino.bsky.social falou na live de hj que a declaração da defesa do Tércio Arnaud era puro chatgpt
joguei no gptzero e kkkkk
(sim, sei que tem falsos positivos, mas nesse caso fodasse né)
joguei no gptzero e kkkkk
(sim, sei que tem falsos positivos, mas nesse caso fodasse né)
Les deux critères majueurs, gptzero appelle ca "perplexity" et "burstiness". J'ai pas les noms exacts en français, je sais même pas si c'est traduit, mais ca va être "perplexité" et "effet de salve" ici.
La "perplexité" ça va mesurer à quel point GPTzero, qui est un modèle de langage, va être
La "perplexité" ça va mesurer à quel point GPTzero, qui est un modèle de langage, va être
February 13, 2025 at 11:53 PM
Les deux critères majueurs, gptzero appelle ca "perplexity" et "burstiness". J'ai pas les noms exacts en français, je sais même pas si c'est traduit, mais ca va être "perplexité" et "effet de salve" ici.
La "perplexité" ça va mesurer à quel point GPTzero, qui est un modèle de langage, va être
La "perplexité" ça va mesurer à quel point GPTzero, qui est un modèle de langage, va être
Thélyson Orélien accusé d’avoir utilisé l’IA : comment fonctionnent les détecteurs comme Pangram, Lucide ou GPTZero
👉 https://l.frandroid.com/WHF
👉 https://l.frandroid.com/WHF
September 23, 2026 at 7:05 AM
Thélyson Orélien accusé d’avoir utilisé l’IA : comment fonctionnent les détecteurs comme Pangram, Lucide ou GPTZero
👉 https://l.frandroid.com/WHF
👉 https://l.frandroid.com/WHF
hilarious, absolutely zero notes.
December 10, 2025 at 7:01 PM
hilarious, absolutely zero notes.
(“work” from a group member) here we fucking go again
June 15, 2026 at 4:01 PM
(“work” from a group member) here we fucking go again
When I fed their own email into their AI detection tool on their website, this is what came out. It's comedy gold.
April 10, 2025 at 5:44 PM
When I fed their own email into their AI detection tool on their website, this is what came out. It's comedy gold.
Hier ist ein Test zu drei KI-Detektoren. Sie liefern sowohl False Positives als auch False Negatives. Wenn Texte manuell überarbeitet werden, ist eine Erkennung kaum noch möglich. Erwischt werden also eher faule Menschen, die keine Ahnung haben.
hub.paper-checker.com/blog/ai-dete...
hub.paper-checker.com/blog/ai-dete...
GPTZero vs Turnitin vs Copyleaks: AI Detector Accuracy Comparison (2026) - Paper Checker
Compare GPTZero, Turnitin, Originality.ai, and Copyleaks accuracy, false positives, pricing, and ESL bias. Data-driven guide for students.
hub.paper-checker.com
June 16, 2026 at 7:23 PM
Hier ist ein Test zu drei KI-Detektoren. Sie liefern sowohl False Positives als auch False Negatives. Wenn Texte manuell überarbeitet werden, ist eine Erkennung kaum noch möglich. Erwischt werden also eher faule Menschen, die keine Ahnung haben.
hub.paper-checker.com/blog/ai-dete...
hub.paper-checker.com/blog/ai-dete...
Need to repeatedly expose myself to formulaic AI-generared outputs in order to turn myself into GPTZero
Y'see it's important to constantly use AI, so you can develop a keen sense of when other people are using AI.
April 30, 2026 at 5:53 PM
Need to repeatedly expose myself to formulaic AI-generared outputs in order to turn myself into GPTZero
Just ran a few of his tweets through gptzero just to confirm my theory, lo and behold:
April 27, 2026 at 10:11 PM
Just ran a few of his tweets through gptzero just to confirm my theory, lo and behold:
Self-billed as ‘the leading AI detector’, GPTZero claims to determine whether a piece of writing was written by a human or LLM.
But GPTZero is pretty much useless for AI detection, and furthermore, its outputs might even be considered harmful.
jamescosullivan.substack.com/p/gptzero-is...
But GPTZero is pretty much useless for AI detection, and furthermore, its outputs might even be considered harmful.
jamescosullivan.substack.com/p/gptzero-is...
GPTZero is useless
If you’ve spent any time at all looking into AI detection then you’ve probably encountered GPTZero.
jamescosullivan.substack.com
June 25, 2025 at 3:14 PM
Self-billed as ‘the leading AI detector’, GPTZero claims to determine whether a piece of writing was written by a human or LLM.
But GPTZero is pretty much useless for AI detection, and furthermore, its outputs might even be considered harmful.
jamescosullivan.substack.com/p/gptzero-is...
But GPTZero is pretty much useless for AI detection, and furthermore, its outputs might even be considered harmful.
jamescosullivan.substack.com/p/gptzero-is...
In what is amusing, this startup called GPTZero (aiming to detect LLM-generated content) reached out along the lines of "we scanned your articles and found no evidence of AI generated content."
I replied: "Yes... duh?"
Surprised paying newsletters would do it.
www.wired.com/story/substa...
I replied: "Yes... duh?"
Surprised paying newsletters would do it.
www.wired.com/story/substa...
November 21, 2024 at 9:02 AM
In what is amusing, this startup called GPTZero (aiming to detect LLM-generated content) reached out along the lines of "we scanned your articles and found no evidence of AI generated content."
I replied: "Yes... duh?"
Surprised paying newsletters would do it.
www.wired.com/story/substa...
I replied: "Yes... duh?"
Surprised paying newsletters would do it.
www.wired.com/story/substa...