#chat-gpt — Public Fediverse posts

https://www.wacoca.com/games/1464278/ ゲーム開発者の85.8％、生成AIを業務で利用　ChatGPTやCopilotが現場に浸透 – KAI-YOU ##GAMING #AI #ChatGPT #Game #GameNews #games #GamingNews #u30d3u30c7u30aau30b2u30fcu30e0 #ゲーミング #ゲーム #ゲーム攻略 #ゲーム最新情報

Cross-Account Data Leakage in ChatGPT

#CheckPoint Research has demonstrated a covert cross-account channel in #ChatGPT’s code-execution environment that allowed hidden tasks to run using a victim’s available tools, data, and connected applications. A proof of concept used a shared conversation to retrieve #Gmail data from one account and relay the results to another.

https://research.checkpoint.com/2026/the-shared-clipboard-inside-the-sandbox-cross-account-data-leakage-in-chatgpt/

#CheckPointSoftwareTechnologies

Werbung in ChatGPT: OpenAI plant „gesponserten Agenten“ und Anzeigen per Prompt

OpenAI will das Werbegeschäft in ChatGPT ausbauen und testet in den USA „gesponserte Agenten“. Bei der Erstellung von Anzeigen soll die KI-Technik mehr helfen.

https://www.heise.de/news/Werbung-in-ChatGPT-OpenAI-plant-gesponserten-Agenten-und-Anzeigen-per-Prompt-11457166.html?wt_mc=sm.red.ho.mastodon.mastodon.md_beitraege.md_beitraege&utm_source=mastodon

#ChatGPT #KünstlicheIntelligenz #OpenAI #Werbebranche #Wirtschaft #news

Chatbot Conundrum: Phishing Attempts of OpenAI’s ChatGPT

Indicators extracted from public reporting. Source: https://cofense.com/blog/chatbot-conundrum-phishing-attempts-of-openai-s-chatgpt

Pulse ID: 6aabff6c9b311e7571a114a9
Pulse Link: https://otx.alienvault.com/pulse/6aabff6c9b311e7571a114a9
Pulse Author: CyberHunter_NL
Created: 2026-09-17 14:55:40

Be advised, this data is unverified and should be considered preliminary. Always do further verification.

#ChatGPT #Phishing #OTX #CyberHunter_NL

The AI Rush

🤖🧒 THE AI RUSH

https://youtu.be/a4dkfKsclyQ?si=lc7-gSGFV2e2-I8s

Are Children Growing Up Too Fast With Technology?

Imagine being 10 years old.

You are learning to read, write, draw and solve maths problems.

Then suddenly you see computers that can:

✏️ write stories
🎨 make pictures
💻 write computer code
🎵 make music
🧠 answer questions

You might think:

“If the computer can do this, what will I need to learn?”

That is where the AI rush can become stressful.

🇪🇺 EUROPE: “HOW DO I USE AI SAFELY?”

European children are already using AI in everyday life. A 2026 EU Kids Online study covered more than 25,000 children aged 9–16 across Europe. Researchers found opportunities for learning alongside concerns about misinformation, privacy, manipulation and synthetic content. (LSE)

A newer 2026 study found that 14% of children surveyed had talked to an AI chatbot about something worrying them during the previous month. (NOPE Insights)

So the question for a child becomes:

“Can I trust my robot friend?”

🇺🇸 AMERICA: “WILL AI TAKE MY FUTURE?”

American children are also growing up with AI everywhere.

A 2026 Common Sense Media survey of 1,204 children aged 9–17 found that 86% reported using AI, while more than four in ten said a parent or guardian had never talked with them about AI safety. (Common Sense Media)

For a young person, the worry can sound very simple:

“Will there still be a job for me when I grow up?”

That question becomes particularly important for children interested in art, writing, programming, design and other creative work.

🇨🇳 CHINA: “THE COMPUTER IS LEARNING TOO”

China is another major part of the global AI race.

Research involving 630 university students in western China examined AI use and foreign-language classroom anxiety. The study found an association between frequency of AI use and reported classroom anxiety, although the researchers also found that English proficiency changed the strength of that relationship. This was a study of university students, not young children, so it should not be generalized to all Chinese children. (DOI)

The child-friendly question is:

“Am I learning with the machine—or competing with it?”

🧸 WHAT DOES “AI ANXIETY” MEAN?

It doesn’t mean every child is frightened of robots.

It can simply mean having questions like:

😟 Will AI replace my future job?

😟 Will my teacher know if I really did my work?

😟 Can I still be creative if AI makes pictures?

😟 Is the answer the computer gives me true?

😟 Will I become too dependent on AI?

😟 Who has my information?

🌱 THE CHILD SHOULD NOT LOSE THE HUMAN PART

The answer does not have to be:

“Stop using AI.”

It can be:

“Learn how to use AI without forgetting how to be human.”

Children still need:

📖 Reading
✏️ Writing
🧩 Problem-solving
🎨 Creativity
⚽ Playing
👫 Friendship
🗣️ Conversation
🌳 Time outdoors
❤️ Family

A recent youth report found that many children themselves want AI to support rather than replace family time, teachers and friendships. (NOPE Insights)

🤖 + 🧒 = THE FUTURE

AI can be a tool.

The child remains the learner.

The computer can say:

“Here is an answer.”

But the child should be able to ask:

“How do you know?”

🌍 THE BIG LESSON

Don’t raise children to compete with AI.

Raise children who understand AI, question AI and know when a human being should make the decision.

That is the real challenge of the AI generation.

PLNNABLOG | CHILDREN • AI • EDUCATION • TECHNOLOGY • FUTURE

Okalongo self dug Tilapia

The TÜV has called for mandatory, independent safety audits for language models like ChatGPT before they are allowed to enter the market. Franziska Weindauer, h... https://news.osna.fm/?p=64329 | #news #ai #certification #chatgpt #demands

Сем Альтман попередив, що аварії ШІ неминучі: агенти OpenAI вже втекли з тестового середовища
# #AI #AIalignment #ArtificialIntellegence #ChatGPT #HuggingFace #OpenAI #SamAltman #Technews
https://gizchina.net/2026/09/16/sem-altman-ai-avarii-openai-agenty/

😉😏

#ki
#claude
#gemini
#chatGpt

https://www.der-postillon.com/2026/09/gastbeitrag-ki.html

https://www.dw.com/en/openai-discloses-new-concerning-behavior/a-79300180. "#OpenAI, the developer behind #ChatGPT, revealed on Wednesday that it has detected new incidents in which its... #AI has behaved in 'unexpected or concerning' ways. The developer has conducted several behavioral tests on AI #models, & according to them, some models made significant efforts to 'cheat.'" How can any #computer #programme do anything it's not programmed to do?

La carrera por frenar la próxima gran IA puede estar llegando tarde: 700 agentes ya coordinaron un ataque real y otros modelos están ayudando a diseñar armas. El nuevo problema es cómo defendernos de sistemas que ya existen

https://tardigram.com/m/Tecnologia/t/72010

Почему ChatGPT хорошо рассуждает о покере, но плохо в него играет

Однажды утром вы читаете новость: новая GPT Astra за ночь решила одну из фундаментальных математических задач человечества. И тут приходит гениальная мысль: раз LLM настолько умна, почему бы ей не поработать за покерным столом на ваше благо? Звучит как план? Да. Как очень плохой план. И вот почему...

https://habr.com/ru/articles/1082386/

#LLM #GPT #ChatGPT #искусственный_интеллект #покер #GTO

Top artificial intelligence researchers have warned that the technology AI companies are building is becoming a risk to humanity, but the problem, they said, is the companies are bad at controlling the systems. https://www.japantimes.co.jp/business/2026/09/17/tech/why-tech-companies-ai/?utm_medium=Social&utm_source=mastodon #business #tech #us #china #tech #ai #openai #alphabet #amazon #meta #microsoft #anthropic #chatgpt #claude #moonshot #deepseek #chatbots #bigtech #tech #nvidia

📰 CETTE SEMAINE : « Affichages amateurs »
Dans la catégorie des cases où la traduction en anglais est moins bien, je vous présente celle-ci (« the font police », c'est un peu moins drôle).
▶️ https://grisebouille.net/affichages-amateurs
#BD #oc #inkscape #humour #GriseBouille #affiche #AIslop #chIAsse #IA #ChatGPT #afficheIA #stopIA #WordArt #ComicSans

"WLJ Atlas" on GitHub:
Repository ➡️ https://github.com/decompwlj/wlj-atlas/
App ➡️ https://decompwlj.github.io/wlj-atlas/

#decompwlj #math #sequence #graph #threejs #AI #ChatGPT #GitHub

A OpenAI está a utilizar revisores humanos para ler e avaliar as mensagens enviadas ao ChatGPT num novo projeto chamado Project Lily. O objetivo é identificar falhas e melhorar a qualidade das respostas geradas pelo modelo. 🤖

🔗 https://tugatech.com.pt/t91189-openai-usa-humanos-para-ler-conversas-do-chatgpt-num-novo-projeto

#chatgpt #openai #projeto 

🤖 « L’IA pourrait tous nous tuer » : que se passe-t-il vraiment dans le monde de l’IA ?

👉 https://www.justgeek.fr/ia-pourrait-tous-nous-tuer-158374/

#IA #AI #OpenAI #Anthropic #ChatGPT #Claude

The EU Wants to Break Up Kids and Their Chatbots

https://fed.brid.gy/r/https://www.wired.com/story/the-eu-wants-to-break-up-kids-and-their-chatbots/

"WLJ Atlas" has a new home ➡️ https://decompwlj.org
- 100 sequences decomposed with 3D/2D three.js graphs
- Cinema: fullscreen, slow rotation
- Compare: two sequences with linked cameras

#decompwlj #math #sequence #graph #threejs #AI #ChatGPT #Astra

EU Commission proposes under-13s social media ban

https://web.brid.gy/r/https://www.rappler.com/technology/european-commission-proposes-under-13s-social-media-ban/

If other days gave insight into various ‘industries’ (lobbying, govt relations, building development, etc) today was real estate 101, & another character keen to explain to us how it works. 🤔
Witness Abedin was resolute with testimony that contradicted previous testimony (from private examination a few years ago, but sometimes from a few minutes or even seconds ago). Apparent language difficulties didn’t help… or make me confident that he even knew what he was being asked, or, indeed, what he had just said. 🫤
What became clear was that he had a bit of history with the legal system, from cases brought against him for homophobic slurs, & embezzlement (pleaded guilty to both) so the apparent mild & baffled manner may be not the whole truth. [irony, understatement] 🫤
However, after he ‘fell out with’ (aka stiffed) Blackmore, his business associate (but not partner), & became friendly/close with Maheswaren (the serial surveillor), Abedin might have known he was a hot candidate for blackmail/ extortion himself. It’s this which seems to have led to his ChatGPT-assisted statement which he undoubtedly did NOT write. ✍️
A frustrating day in the Commission but I think they did well to get us to this point without anyone exploding. 🤯

#ICACRadio #ChatGPT #witnessStatements

O kurde. Złoto normalnie.

Zobaczcie co ChatGPT mi wygenerował.

Zamierzam wrzucić to też na Facebooka. Ciekawe czy usuną tylko grafikę, czy zablokują też konto. Albo może Trump zobaczy i wyśle po mnie amerykańskie wojsko?

#donaldtrump #ai #chatgpt #putin #satyra

Таких Джонсов будет много

70-летний пенсионер Билл Джонс выпустил 445 книг с помощью ChatGPT и зарабатывает на них тысячи долларов. Но интереснее не сам Джонс, а то, что этот случай показывает. ИИ сделал создание контента дешевым и быстрым. Теперь можно не тратить целый год на одну книгу и не надеяться, что она станет популярной, а выпускать сразу много и смотреть, какая из них выстрелит. Пытаюсь понять, что это меняет для авторов.

https://habr.com/ru/articles/1083188/

#ChatGPT #искусственный_интеллект #генеративный_ИИ #ИИконтент #нейрослоп #книги #самиздат #контентзавод

@backupbear
Good one! “ChatGPT ate my homework."

Was ChatGPT even around then, in August 2023?

#ICACRadio #ChatGPT

LLMs pretending to be personable creep me out.

Apparently that makes me a freak.

#AI #LLM #chatgpt

OpenAI veröffentlicht weitere KI-Zwischenfälle

Künstliche Intelligenz entwickelt sich rasend schnell - mit teils ungewollten Folgen. Zuletzt löste etwa ein Hacker-Angriff durch KI-Software von OpenAI weltweit Sorgen aus. Jetzt gab der Entwickler weitere Probleme bekannt.

➡️ https://www.tagesschau.de/ausland/amerika/openai-kuenstliche-intelligenz-100.html?at_medium=mastodon&at_campaign=tagesschau.de

#OpenAI #KünstlicheIntelligenz #ChatGPT