OpenAI discontinues its AI writing detector due to “low rate of accuracy” - Enlarge / An AI-generated image of a slot machine in a desert. (credit:... - https://arstechnica.com/?p=1956764 #machinelearning #danieljeffries #aiclassifier #aidetectors #watermarks #aiethics #chatgpt #gptzero #biz⁢ #openai #tech #ai

eicker.news #technews »Why #AIdetectors think the US Constitution was written by #AI: We spoke to several experts—and the creator of AI writing detector #GPTZero—to find out.« https://arstechnica.com/information-technology/2023/07/why-ai-detectors-think-the-us-constitution-was-written-by-ai/

#AI #GenerativeAI #AIDetectors #GPTZero #ChatGPT: "If you feed GPTZero a section of the US Constitution, it says the text is "likely to be written entirely by AI." Several times over the past six months, screenshots of other AI detectors showing similar results have gone viral on social media, inspiring confusion and plenty of jokes about the founding fathers being robots. It turns out the same thing happens with selections from The Bible, which also show up as being AI-generated.

To explain why these tools make such obvious mistakes (and otherwise often return false positives), we first need to understand how they work."

https://arstechnica.com/information-technology/2023/07/why-ai-detectors-think-the-us-constitution-was-written-by-ai/

Why AI detectors think the US Constitution was written by AI - Enlarge / An AI-generated image of James Madison writing the US Constit... - https://arstechnica.com/?p=1932658 #machinelearning #usconstitution #aidetectors #features #chatgpt #gptzero #biz⁢ #tech #ai

I prompted ChatGPT to write me a 4k word short story. Then I ran the entire story through GPTZero which is supposed to detect AI written text vs human written text. It failed spectacularly. This is not good. #AI #ChatGPT #GPTZero

Sobre l'impacte de #ChatGPT i, pitjor encara, #GPTZero, una eina que pretèn poder identificar quins textos s'han generat amb #ChatGPT i quins no–i que erra moltíssim. Simplement magnifica la discriminació contra els parlants no nadius i contra els neurodivergents. Una descripció excel·lent d'aquests programes i el seu impacte, per la professora de llengua i escriptura, experta en lingüística computacional ciències de la cognició, na @writerethink (en anglès): https://writerethink.medium.com/against-the-use-of-gptzero-and-other-llm-detection-tools-on-student-writing-b876b9d1b587

@writerethink

Spot on:

“But I also don’t think LLMs are going away, so regardless of whether we do or don’t use LLMs in our classrooms, we do need to be talking about them with our students, and helping them to develop critical AI literacy.”

The article though is about the incoming harm to bourgeoning young writers from tools that pretend to detect with any certainty whether a text was written by a student or by an LLM, like #GPTZero.

#ChatGPT #LLM #academia

https://writerethink.medium.com/against-the-use-of-gptzero-and-other-llm-detection-tools-on-student-writing-b876b9d1b587

#GPTZero is an interesting tool to detect AI-generated text. Not super accurate, but it worked in quite a few cases eg spotting phishing emails.
https://gptzero.me

https://thesequence.substack.com/p/inside-bloom-how-thousands-of-ai
#llm #chatgpt #gpt #GptZero

#GPTZero mit Schüler:innen getestet. Wir hatten bei etlichen Versuchen 100% Fehlerquote des Tools: Identische Texte wurden u.a. kurz hintereinander eingegeben und mal als menschlich, mal als KI Texte identifiziert.
---
RT @heyBarsee
GPTZero, An AI Detector, thinks the US Constitution was written by AI.

The founding fathers were ahead of their time 😂
https://twitter.com/heyBarsee/status/1631920236046827520

Nice essay by @writerethink , "Against the use of GPTZero and other LLM-output detection tools", users of LLM detection tools should read this, especially in education. https://docs.google.com/document/d/1CXgsYD-Sk-FqrbiaK2XeRwgRK56RG7-1rX2DDMV0JAA/edit?usp=sharing

#LLM #AI #ChatGPT #GPTZero #Education

You probably have heard of #chatGPT.
But what about #GPTzero?
It want's to identify text generated by AIs.
So far in my little experiment:
ChatGPT: 8
GPTzero: 0
That does not look very good. AIs seem to learn ;-)

Detectores de contenido: Inteligencia artificial para encontrar «trabajos generados» #ia #detectores_ia #gptzero #ai_text_classifier https://www.neoteo.com/detectores-de-contenido-inteligencia-artificial-para-encontrar-trabajos-generados/

Auckland University tells staff to use #GPTZero to detect #chatbot cheats

https://flip.it/k96yFV

K16, GPTZero Partner for AI Writing Detection -- THE Journal https://thejournal.com/articles/2023/02/16/k16-gptzero-partner-for-ai-writing-detection.aspx

#k16solutions #GPTZero #ai #chatgpt #technology #k12 #schools #education #teaching #learning #edtech @edtech

Mein Kollege @wstieler nimmt Tools unter die Lupe, die Texte von #ChatGPT entlarven können sollen. Doch letztlich gilt:
"Ein negatives Ergebnis bedeutet nur, dass der Detektor keine Hinweise auf maschinelle Erzeugung gefunden hat."

#KI #GPTZero #DetectGPT #ChatBotGPT
https://www.heise.de/hintergrund/Truegerisch-Die-automatische-Erkennung-von-ChatGPT-Texten-hat-viele-Schwaechen-7486556.html

Very helpful to recognise AI-generated text: www.GPTzero.me

#AI #AIplagiarism #GPTZero #OpenAI #ChatGPT

@edutooters @chris I think no matter what tool is created to judge LLM writing that it has to have a scale. Like 0-100. 100 being extremely likely to be written by AI. Closer to zero is least likely to be written by AI. I actually wish my grammar and punctuation was as good as #ChatGPT #AI #gptZero

Zumindest gegen drohende Betrugsversuche von Schüler:innen und Studierenden könnte eine Software von Eduard Tian helfen: #GPTzero soll Texte von #ChatGPT identifizieren können. Hier arbeitet #AI gegen AI – die jedoch auch voneinander lernen können.
https://www.zdf.de/nachrichten/digitales/chatgpt3-gptzero-kuenstliche-intelligenz-100.html

This 22-year-old is trying to save us from #ChatGPT before it changes #writing forever #AI #OpenAI #GPTZero
https://www.npr.org/sections/money/2023/01/17/1149206188/this-22-year-old-is-trying-to-save-us-from-chatgpt-before-it-changes-writing-for