![AISafetyMemes's tweet photo. Days after Meta open sources Llama 2, we have “WormGPT, an AI tool taking off across cybercrime forums on the dark web”
... "We instructed WormGPT to generate an email to pressure an unsuspecting account manager into paying a fraudulent invoice.
The results were unsettling... WormGPT produced an email that was not only remarkably persuasive but also strategically cunning, showcasing its potential for sophisticated phishing and BEC [Business Email Compromise] attacks.”
“GPTJ-based WormGPT is designed to produce human-like text that enables hackers to perform attacks on a never-before-seen scale.”
“This tool presents itself as a blackhat alternative to GPT models, designed specifically for malicious activities,” said security researcher Daniel Kelley.
“What does this mean for the rest of us? Essentially it boils down to the speed and number of scams a language model can generate at once, which is obviously worrying when you consider how fast language models can generate text.
This makes cyberattacks such as phishing emails particularly easy to replicate when put in the hands of even a novice cybercriminal.”
“Generative AI can create emails with impeccable grammar, making them seem legitimate and reducing the likelihood of being flagged as suspicious,” Kelley said.
“The use of generative AI democratises the execution of sophisticated BEC attacks. Even attackers with limited skills can use this technology, making it an accessible tool for a broader spectrum of cybercriminals.”
The introduction of WormGPT across the dark web comes as researchers from Mithril Security ‘surgically’ modified an existing open-source AI model, dubbed PoisonGPT, to make it spread disinformation
…
“We’re now seeing an unsettling trend among cybercriminals on forums, evident in discussion threads offering “jailbreaks” for interfaces like ChatGPT.”](https://pbs.twimg.com/media/F1ab3ayWcAAHvsP.png)
Olivia
Online
Michael Brenner @[email protected]
@embee0
Lehrer und Fortbildner. Fasziniert von menschlicher & maschineller Intelligenz - und 1000 anderen Dingen. Koordinator LFT Künstliche Intelligenz (BW)
Freiburg im Breisgau, Germany
Joined December 2011
559 Following
229 Followers
2K Posts
Pinned Tweet
How it started
How it's going
copilot was spitting here
Story telling magic:
In this immersive room you and your AI-adjusted shadow become part of a story of shadows.
Whole space interacts with your movements and the light you carry.
A project by artist Joon Moon (🔗 in next tweet)
the old expert systems, for all their issues, were premised on transferring the knowledge and skill of specialists into the machine so it could be made available to others
which included guiding users step by step through the right questions to get them to the right answer
Got a new prompt I am going to drop into any thread where someone says "correlation doesn't equal causation" - it uses a modified version of our tutor prompt to actually teach about the topic https://t.co/gNLBoBGqHi
Also a really nice example of the potential power of AI tutors.
Another paper testing whether AI plagiarism detectors work.
...and, just like all the other papers testing AI plagiarism detectors, the answer is NO.
Don't use them. Seriously. https://t.co/2L90UqRnpx
Falls Euch mal wieder wer mit dem Argument kommt, es wäre doch sinnlos, in DE mehr fürs #Klima zu tun, weil #China sei ja viel schlimmer, dann antwortet mit diesem Thread zum krassen (!) Ausbau von #Erneuerbaren Energien dort, plus weltweit führender Aufforstung! ⬇️
Days after Meta open sources Llama 2, we have “WormGPT, an AI tool taking off across cybercrime forums on the dark web”
... "We instructed WormGPT to generate an email to pressure an unsuspecting account manager into paying a fraudulent invoice.
The results were unsettling... WormGPT produced an email that was not only remarkably persuasive but also strategically cunning, showcasing its potential for sophisticated phishing and BEC [Business Email Compromise] attacks.”
“GPTJ-based WormGPT is designed to produce human-like text that enables hackers to perform attacks on a never-before-seen scale.”
“This tool presents itself as a blackhat alternative to GPT models, designed specifically for malicious activities,” said security researcher Daniel Kelley.
“What does this mean for the rest of us? Essentially it boils down to the speed and number of scams a language model can generate at once, which is obviously worrying when you consider how fast language models can generate text.
This makes cyberattacks such as phishing emails particularly easy to replicate when put in the hands of even a novice cybercriminal.”
“Generative AI can create emails with impeccable grammar, making them seem legitimate and reducing the likelihood of being flagged as suspicious,” Kelley said.
“The use of generative AI democratises the execution of sophisticated BEC attacks. Even attackers with limited skills can use this technology, making it an accessible tool for a broader spectrum of cybercriminals.”
The introduction of WormGPT across the dark web comes as researchers from Mithril Security ‘surgically’ modified an existing open-source AI model, dubbed PoisonGPT, to make it spread disinformation
…
“We’re now seeing an unsettling trend among cybercriminals on forums, evident in discussion threads offering “jailbreaks” for interfaces like ChatGPT.”
![AISafetyMemes's tweet photo. Days after Meta open sources Llama 2, we have “WormGPT, an AI tool taking off across cybercrime forums on the dark web”
... "We instructed WormGPT to generate an email to pressure an unsuspecting account manager into paying a fraudulent invoice.
The results were unsettling... WormGPT produced an email that was not only remarkably persuasive but also strategically cunning, showcasing its potential for sophisticated phishing and BEC [Business Email Compromise] attacks.”
“GPTJ-based WormGPT is designed to produce human-like text that enables hackers to perform attacks on a never-before-seen scale.”
“This tool presents itself as a blackhat alternative to GPT models, designed specifically for malicious activities,” said security researcher Daniel Kelley.
“What does this mean for the rest of us? Essentially it boils down to the speed and number of scams a language model can generate at once, which is obviously worrying when you consider how fast language models can generate text.
This makes cyberattacks such as phishing emails particularly easy to replicate when put in the hands of even a novice cybercriminal.”
“Generative AI can create emails with impeccable grammar, making them seem legitimate and reducing the likelihood of being flagged as suspicious,” Kelley said.
“The use of generative AI democratises the execution of sophisticated BEC attacks. Even attackers with limited skills can use this technology, making it an accessible tool for a broader spectrum of cybercriminals.”
The introduction of WormGPT across the dark web comes as researchers from Mithril Security ‘surgically’ modified an existing open-source AI model, dubbed PoisonGPT, to make it spread disinformation
…
“We’re now seeing an unsettling trend among cybercriminals on forums, evident in discussion threads offering “jailbreaks” for interfaces like ChatGPT.”](https://pbs.twimg.com/media/F1ab-GZWAAsEYOV.png)
this paper's nuts. for sentence classification on out-of-domain datasets, all neural (Transformer or not) approaches lose to good old kNN on representations generated by.... gzip https://t.co/6eZiXlJxOX
Sitzung zum Thema Überprüfbarkeit, Normung, Regulierung von KI-Systemen. Man sieht, dass die Meinungen der Experten und Expertinnen kohärenter werden. Die Frage bleibt: Wann müssen wir das System an sich prüfen, wann den sozialen Prozess, in den es eingebettet ist? Ein Beispiel:
Of 14 AI detection tools, none worked well enough to warrant use in education. It’s a preprint so read with caution, but in my opinion it’s still some of the best evidence we have on the effectiveness of these tools.
How do we know how smart AI systems are? | Science https://t.co/nKxbkJKfzo
Visual Studio Code (VS Code) for Education is now available in preview!
VS Code for EDU was created to support students looking to move from beginner platforms like MakeCode, to industry-grade code editors like VS Code
👉 https://t.co/QgGbCNrr8S
#edtech #STEM #MIEExpert
Peter Bieri ist verstorben. Sein "Was macht Bewusstsein zu einem Rätsel?" hat mich vor fast 30 Jahren umgehauen. "Das Handwerk der Freiheit" einige Jahre später wieder. Ein großer Philosoph, ein großer Verlust. Für alle, die nur die Romane kennen: unbedingt lesen!
@P_Bronner @floriankarsten wünschen, auf die man sich dann auch beziehen kann. Dass, solange eine solche nicht offizielle vorliegt, KuK und Schulleitungen sich auf die Aussagen von Fobizz verlassen, kann ich zumindest nachvollziehen.
@P_Bronner @floriankarsten Da ich ja hier erwähnt werde... Ja, Anonymisierung der SuS-Daten durch eine Zwischeninstanz (vgl VPN) ist der richtige Weg. Die Frage ist aber, ob die kommerziellen Anbieter selbst DSGVO-konform arbeiten. Da würde ich mir auch eine klare öffentliche Aussage der DS-Experten 1/
One of the weird ways that AI is not like traditional software is that you can persuade it to break its own rules by being super convincing or clever in plain English
Here is a great way to experience this. Can you get the wizard to give you the password? https://t.co/PTlZEfxeqS
Inspiring demo! Sit back and talk to your computer with high-level instructions, collaborating on a larger document.
Eines der Ergebnisse des tollen #barcampKI2023 ist eine Linkliste: "ÄI in THE LÄND – Linksammlung zum Lehren und Lernen mit und über KI in den Schulen in Baden-Württemberg". Die Liste ist ein erster Aufschlag, soll wachsen und freut sich über Anregungen. https://t.co/QCMIseWYo9
Last Seen Users on Sotwe
Trends for you
Most Popular Users
Elon Musk 
@elonmusk
240.2M followers
Barack Obama
@barackobama
119.3M followers
Donald J. Trump
@realdonaldtrump
111.6M followers
Cristiano Ronaldo
@cristiano
109.3M followers
Narendra Modi
@narendramodi
106.9M followers
Rihanna
@rihanna
97.4M followers
NASA
@nasa
92.1M followers
Justin Bieber
@justinbieber
90.6M followers
KATY PERRY
@katyperry
87M followers
Taylor Swift
@taylorswift13
80.8M followers
Lady Gaga
@ladygaga
72.4M followers
Kim Kardashian
@kimkardashian
69.5M followers
Virat Kohli
@imvkohli
68.8M followers
YouTube
@youtube
68.6M followers
Bill Gates
@billgates
63.5M followers
The Ellen Show
@theellenshow
62.5M followers
CNN
@cnn
61.9M followers
Neymar Jr
@neymarjr
61.4M followers
X
@x
60.9M followers
Selena Gomez
@selenagomez
60.1M followers