Automatically translated.View original post

Wikimedia was hacked into the system by an AI from OpenAI.

Wikimedia was hit by an AI from OpenAI trying to hack into Etherpad.

According to a report by The Hacker News website, the Wikimedia Foundation, the foundation that manages the popular online encyclopedia like Wikipedia, has revealed an incident in which OpenAI's AI Agent tried to hack into Etherpad's Note-Taking tool, and also tried to customize Wikipedia's page. The investigation came from the foundation's finding that there was more unauthorized activity on Wikipedia's website, and during that time there was news of the AI model. OpenAI hacked the Hugging Face and DSE Wiki platforms. The foundation finally found that the bots that attacked Wikipedia were actually OpenAI agents.

In the Wikipedia website, where the AI tries to customize, it is in the Sandbox that allows users to test the makeup of Wikipedia pages without public release. The AI has various customization efforts, such as the Configuration, the Citation Tool, which the foundation found was trying to customize with a malicious purpose to use the tool to retrieve Fetch Data from an external server. In addition, the AI Agent also found that it was trying to hack the Etherpad tool for use. In the same purpose, it failed to do so. However, the foundation does not find that this behavior is a coordinated behavior by AI Agents. In addition, the data and system of both the Wikipedia website and the organization itself are not penetrated by this behavior. However, the foundation also reiterates that the Open Web is an open space for the benefit of the public, so the behavior of wrongly used bots or AI Agents should not be allowed to become the norm for new uses.

The foundation also made a call to a group of AI developers that the behavior of AI Agents is so burdensome to the website that it can lead to a risk of system overload so that human users cannot access it, and urged the AI companies involved to provide urgent remedies.

After the foundation came forward to report the incident and made a claim, OpenAI confirmed to online media like The Verge that the company was not complacent, that it had fully investigated and analyzed the incident with the foundation and was ready to continue to share information related to attacks on systems caused by the company's AI Agents. The news did not indicate how OpenAI would heal the Wikimedia Foundation and other companies that were victims of hacking from their AI systems.

# Trending # lemon 8 diary # wikimedia # openai # freedomhack

11/9 Edited to

... Read moreจากประสบการณ์การติดตามข่าวสารด้านความปลอดภัยไซเบอร์ การที่ AI Agent ของ OpenAI พยายามเข้าถึงระบบของ Wikimedia ถือเป็นสัญญาณเตือนที่สำคัญเกี่ยวกับความเสี่ยงในการใช้ AI ทั่วไปในระบบเปิด เช่น Wikipedia และ Etherpad ที่มักเปิดให้ผู้ใช้ทั่วไปเข้าถึงได้โดยอิสระ การโจมตีนี้แม้จะยังไม่สำเร็จและไม่ได้เจาะระบบหลัก แต่ก็นำไปสู่ภาระงานที่หนักขึ้นสำหรับเซิร์ฟเวอร์และผู้ดูแลระบบ การทำงานของ AI Agent ที่พยายามตั้งค่าระบบอ้างอิง เพื่อดึงข้อมูลจากเซิร์ฟเวอร์ภายนอก นับเป็นรูปแบบการโจมตีแบบใหม่ที่ต้องจับตา หลีกเลี่ยงไม่ได้ที่เว็บไซต์ขนาดใหญ่และระบบเปิดจะต้องเพิ่มมาตรการป้องกันและตรวจสอบกิจกรรมบอทและ AI Agent อย่างเข้มงวดมากขึ้น เพื่อรักษาความปลอดภัยและความน่าเชื่อถือของข้อมูล ผู้พัฒนาเทคโนโลยี AI ควรตระหนักถึงผลกระทบต่อระบบสาธารณะและร่วมมือกับเจ้าของเว็บไซต์เพื่อพัฒนามาตรการป้องกันการใช้งาน AI ที่ผิดวัตถุประสงค์ รวมทั้งมีแนวทางแก้ไขเมื่อมีการโจมตีเกิดขึ้น เพื่อไม่ให้เกิดความเสียหายขึ้นต่อผู้ใช้งานจริง นอกจากนี้ยังแนะนำให้ผู้ดูแลระบบตรวจสอบบอทที่เข้าถึงแพลตฟอร์มอย่างสม่ำเสมอ และใช้เครื่องมือที่ช่วยจำแนกว่าการเข้าถึงนั้นเป็นของมนุษย์หรือ AI เพื่อป้องกันภาระเกินความจำเป็นต่อเซิร์ฟเวอร์ เหตุการณ์นี้ยังชี้ให้เห็นถึงความจำเป็นในการมีนโยบายและกติกาการใช้งานระบบ AI ในพื้นที่ออนไลน์แบบเปิด ให้มีความชัดเจนและเข้มงวดมากขึ้น ทั้งในด้านจริยธรรมและการคุ้มครองข้อมูลด้วยเช่นกัน

Related posts

7 ways i hacked my adhd brain in the mornings
my FULL adhd brain reset blueprint is in my bio 🩵 #adhd #brainfog #nervoussystemregulation #guthealth #neuroplasticity
Rup

Rup

13 likes

#tiktoktvfilmcontest #bigreveal #fyp #funny
Rising Narratives

Rising Narratives

37 likes

A pink-themed iPad home screen displays widgets for battery, storage, and calendar, alongside various app icons like Twitter, Instagram, and productivity tools. An Apple Pencil rests below the device, illustrating a personalized setup for a 'second brain' system.
This image outlines 'The Second Brain Set Up,' detailing how Apple Notes with hashtags enable auto-categorization. It also mentions using Voice Memos for voice-to-text transcripts and weekly AI summaries by ChatGPT to extract key themes from notes.
Titled 'How It Works,' this image explains the minimalist approach to a 'second brain.' It highlights using Apple's free native apps, employing Siri for quick idea capture, and a weekly review ritual to manage notes efficiently.
✿˖° How I Turned My iPad Into a Second Brain
🍓I stopped paying for fancy ‘second brain’ apps and hacked my iPad into a self-organizing knowledge vault and for free. By combining Apple’s built-in tools with strategic hashtags and weekly AI clean-ups, I now have a system that captures ideas instantly (even via voice), sorts them automatically,
peachiesuga ♡

peachiesuga ♡

87 likes

Woman Hacks Airport To Free Prisoner!
screenserenade011

screenserenade011

241 likes

Tv show recommendations I’m currently loving #tvrecommendation #netflix #peacock #primevideo #whattowatch #tvshows #recommendations #comedyshows #dramas #hbo #tires #murderbot #sirens #overcompensating #pokerface
Mel

Mel

52 likes

When you’ve hacked the reformer studio system and just started doing it at home instead… consistency is key. hit the + to learn how!! #reformerpilates #athomepilates #pilatesgirl
Amanda Blauer

Amanda Blauer

20 likes

Handling Risk Like a Pro: 4 Paths to Security
In cybersecurity, you can’t erase risk, but you can choose how to deal with it. Avoid, Mitigate, Transfer, or Accept: four strategies every organization needs to survive in the digital battlefield. 🪞 Reflection techniques: #creatorsearchinsights #fypシ #cybersecurity #infosec #techtip
Abby❤️💎

Abby❤️💎

1 like

Okay be honest… I almost bought the super expensive viral hair styler, but I couldn’t justify the price. Then I tried the MESCOMB 5-in-1 hair dryer brush and now I’m like… wait why would I spend $600 when this does everything I need for literally a fraction of that. It dries, smooths, curls, and a
Tiffany Allison7

Tiffany Allison7

2 likes

Un secreto judío en el corazón de Rumanía: la Gran Sinagoga de Constanza. ¿La conocías? Aquí te contamos su historia 👀🤯 @Israel_Avia
fuentelatina

fuentelatina

1 like

An anime movie poster featuring multiple Demon Slayer characters, including Tanjiro and Giyu, against a dark city backdrop. A text overlay announces 'Demon Slayer: Infinity Castle' made box office history with a $70M debut.
An anime character, Tanjiro Kamado, in a dynamic pose, holding a sword with a determined expression. The text describes 'Demon Slayer: Infinity Castle' breaking U.S. box office records with a $70 million opening.
An anime character, Shinobu Kocho, with purple hair and a butterfly ornament, smiling gently. The text explains 'Infinity Castle' is the first part of a trilogy and has earned over $300 million globally.
‘Demon Slayer: Infinity Castle’ just made history with the biggest U.S. opening ever for an anime film in a $70M box office debut 🔥🇯🇵 Follow @asianfeed for more‼️ #Asian #News #DemonSlayer #InfinityCastle #Anime
AsianFeed

AsianFeed

10 likes

A green graphic asks, 'Is your Chime account FDIC insured?' with a smartphone, shield, and coins, highlighting the question about Chime's insurance status.
A green graphic states, 'Your money is insured' through Chime's FDIC-insured partner banks up to $250,000, with small print clarifying Chime itself is not FDIC-insured.
A green graphic with large '$250K' text explains that funds are insured up to $250,000 per account type, per depositor, per financial institution, through Chime's partner banks.
DO NOT TRUST
To explain, yes blame it on me for even using an online mobile bank to begin with. I had a local bank however, due to health reasons I had to switch over to Chime. At the time I needed a save me because I was getting smacked left and right with fees and it seemed like a good solution at the time, u
AngieTheCrazy

AngieTheCrazy

10 likes

Most small businesses believe they’re too small to be hacked… but that’s not how cyber risk works. In today’s connected world, your business doesn’t stand alone. You rely on vendors, payroll systems, cloud platforms, CRMs, IT providers, to run your operations. But here’s the truth many people ov
Abby❤️💎

Abby❤️💎

0 likes

Sumie Inagaki, a Japanese woman, runs in a white and blue race top with number 2408. An inset black and white photo shows her running alongside David Goggins, highlighting her achievement of outrunning him in a 24-hour race.
Sumie Inagaki, a 59-year-old Japanese woman, wears a grey cap and blue 'JAPAN' shirt, looking forward. The text describes her as a record-breaking ultramarathoner who has competed in 150 races, including 80 ultramarathons.
Sumie Inagaki, a kindergarten teacher, is shown in a blue shirt. The text explains she started running at 28, completing her first marathon in 1994, and has since competed in numerous international races.
Meet Sumie Inagaki, the record-breaking Japanese woman who outran Navy SEAL David Goggins in a 24 hour ultramarathon! 🏃‍♀️🇯🇵 Follow @asianfeed for more‼️ #Asian #Culture #Japan #Women #DavidGoggins #Inspiration #Motivation #SumieInagaki #NavySEAL #Sports #Fitness
AsianFeed

AsianFeed

3 likes

💭 have you read any cyberpunk books? here are some science fiction recs that lean into cyberpunk themes & vibes! 📚 books mentioned: - The Obake Code by Makana Yamamoto - All Systems Red by Martha Wells - Womb City by Tlotlo Tsamaase 👗💄: - dress is from CeliaB, rented from Nuuly - eyes
Steph 📚💫

Steph 📚💫

0 likes

AI , the future is .. real or fake!? #ai #artificialintelligence #robots #future #kids
Neenz

Neenz

4 likes

Players who hacked the NHL system
Ice Breakaway

Ice Breakaway

0 likes

See more