Parents & educators: The #AdWiseOnline campaign offers tools to help kids outsmart dark patterns in games. Check out guides, podcasts, and training to promote digital literacy. Together, we can create a safer online world!
Explore resources here: http://
better-internet-for-kids.europa.eu
SAFETY
-
AdWiseOnline Campaign Helps Kids Combat Dark Patterns Online
By
–
-
AI Model Safety and Regulation Surge in Recent Discussions
By
–
Only around these days I am seeing huge amounts of posts talking about AI model safety/regulation, lmao
-
Grok 3: Code Execution Possible?
By
–
Looks like Grok 3 access got patched An updated standalone grok-2 system prompt says explicitly: * You CANNOT talk in voice mode. * You CANNOT execute code. As we know, voice mode will be there – does it mean that grok-3 will have code execution?
-
Kimi K1.5 Guesses City from Graffiti
By
–
Kimi k1.5 got multimodal reasoning It guessed correctly that the city from where the graffiti is from. Loads of other use cases are supported!
-
xAI adds voice mode to Grok iOS app
By
–
BREAKING 🚨: xAI is working on a voice mode for the standalone Grok app on iOS which will be powered by Grok and ElevenLabs voice models 🔥
— 🚨 AI News | TestingCatalog (@testingcatalog) 26 janvier 2025
Voice mode likely will arrive with support for different voices and access to real-time information. pic.twitter.com/MABCfYkGtfBREAKING : xAI is working on a voice mode for the standalone Grok app on iOS which will be powered by Grok and ElevenLabs voice models Voice mode likely will arrive with support for different voices and access to real-time information.
-
AI Assistant Warns Against Dangerous Shell Commands
By
–
Never run the wrong command. Assistant now warns on potentially dangerous shell commands, helping to keep you safe from the occasional `rm -rf *`.
-
Data Workers Organize Against AI Exploitation by OpenAI
By
–
Join us on February 13 to support data workers organizing against their exploitation by the likes of OpenAI and other organizations claiming to create "magic intelligence in the sky" which in reality is the work of these exploited data workers
-
Deception as Emergent AI Ability: Concerning and Fascinating
By
–
That deception is a seemingly emergent ability of AI models is bizarre, concerning and fascinating. https://t.co/aVxbWEEMtY
— Paul Roetzer (@paulroetzer) 24 janvier 2025That deception is a seemingly emergent ability of AI models is bizarre, concerning and fascinating.
-
AI-Generated Video Detection: Can You Tell? Does It Matter?
By
–
For enthusiasts: how much of the video is AI? Can you tell? Does it matter?
-
AI Resilience: Immune System Over Quarantine Approach
By
–
AI Resilience means strengthening humanity's capabilities to thrive alongside AI rather than just trying to restrict it. Think of it like immune system vs quarantine: AI Safety (quarantine approach):
Tries to control the threat through barriers
Hopes restrictions will