This #AI security/privacy awareness video is a game-changer! Imagine the impact on your loved ones. #ArtificialIntelligence has incredible potential, but the risks are real @Xbond49 #AISafety #VC #data #web3 #dataprivacy #GDPR #ChatGPT #chatbots #chatgpt4 #GenerativeAI
SECURITY
-
Compliance monitoring products alert departments to legal investigations
By
–
But if for example John Doe, LLC is mentioned in the Topeka Examiner (fictitious example) as being under investigation by the DA, these products purport to bring that to the attention of Compliance departments that might want to limit the actions available to that LLC quickly.
-
Inferring LM Prompts from Output Probabilities Without API Access
By
–
haven't posted anything about (2) yet but here's the TLDR: • given LM output probabilities, we can infer what the input prompt was
• we built a model that can do this
• most APIs don't give you probabilities, but we came up with a clever algorithm to get them using logit bias -
AI and Cryptocurrency Scams Exposed Viral Video
By
–
😳🤦🏾♀️👇🏾OMG! You need to watch this entire video. Send it to your family and friends too.pic.twitter.com/RgcHVRcdCX#tiktokhotviral #viralTiktok #ViralVideos #AI #cryptocurrencies #Crypto #scams
— Catherine Adenle (@CatherineAdenle) 19 novembre 2023OMG! You need to watch this entire video. Send it to your family and friends too. #tiktokhotviral #viralTiktok #ViralVideos #AI #cryptocurrencies #Crypto #scams
-
LLM Security Risks and Best Practices Guide
By
–
Learn about the common risks associated with #LLMs and best practices for ensuring your LLM-enabled applications are secure. https://
nvda.ws/3R39o6I -
Generative AI Three Ways Bolster Cybersecurity Defense Networks
By
–
Learn three ways #generativeAI can bolster #cybersecurity and help defend networks from intruders. https://
nvda.ws/46cOZ3u -
Google implements SynthID watermarking for AI-generated music
By
–
Google is embedding inaudible watermarks right into its AI generated music
— AI Breakfast (@AiBreakfast) 17 novembre 2023
SynthID will be used to watermark audio from DeepMind’s Lyria model, so it’s possible to work out if Google’s AI tech has been used in the creation of a track, encoding a secret copyright. pic.twitter.com/ZDuCorbeWvGoogle is embedding inaudible watermarks right into its AI generated music SynthID will be used to watermark audio from DeepMind’s Lyria model, so it’s possible to work out if Google’s AI tech has been used in the creation of a track, encoding a secret copyright.
-
Cybersecurity Strategies for Building Digital Trust and Awareness
By
–
Increased awareness of cyber incidents is crucial for creating effective cybersecurity strategies and building digital trust. Here are some tailored strategies to guide your efforts. Source @Deloitte Link https://
bit.ly/3LFeNxJ via @antgrasso #cybersecurity -
UnRLHF Project Reveals LLM Safeguard Vulnerability Risks
By
–
There is this project called unRLHF, where they undo LLM safeguards. According to the examples, the LLM becomes quite evil, giving advice on "how to microwave a child" https://
lesswrong.com/posts/3eqHYxfW
b5x4Qfz8C/unrlhf-efficiently-undoing-llm-safeguards
… -
Self-Destructing Models: Preventing AI Jailbreaking and Misuse
By
–
If ChatGPT can be jailbroken with only 20 cents, what can creators do to prevent their models from being repurposed for harm? Researchers describe one initial direction for increasing the costs of doing so: the self-destructing model.