Ethical AI Frameworks to help ensure fairness, transparency, and social good in AI. Apply Now at https://
indiaai.gov.in/article/expres
sion-of-interest-for-safe-trusted-ai-projects-under-indiaai-mission
… @GoI_MeitY @abhish18 @_DigitalIndia #EthicsInAI #IndiaAI #ResponsibleAI #AIForGood
SAFETY
-
India AI Mission Launches Ethical AI Framework Projects
By
–
-
Jailbreaking frontier AI models still works in 2025
By
–
A year ago I thought jailbreaks would be fixed by now, or at least chased into newer modalities, but in 2025 you’ll still be able to jailbreak frontier models with a script that spams variations of tLAkIng Liek tHis
-
AI Watermarking Tools for Content Authentication and Traceability
By
–
Watermarking & Labelling tools to track and authenticate AI-generated content, ensuring traceability and security. Apply Now at https://
indiaai.gov.in/article/expres
sion-of-interest-for-safe-trusted-ai-projects-under-indiaai-mission
… @GoI_MeitY @abhish18 @_DigitalIndia #ResponsibleAI #Watermarking #SyntheticMedia #ContentSecurity -
Public video shows interruption during 2016 AI talk
By
–
The video of him interrupting my 2016 talk is also public, and you don’t have to take my word for it that that’s him. He doesn’t mention artificial curiosity, just predictability minimization.
-
AI Is Usually Bad At Math. Here’s Why It Matters
By
–
AI Is Usually Bad At Math. Here’s Why It Matters
#AI #AIio #BigData #ML #NLU #Futureofwork @Kevin_Jackson @sallyeaves @schmarzo @kdnuggets @akwyz @BrettKing @TechNative @MiaD -
Deterministic workflows build trust in AI agents
By
–
deterministic workflows make "trusting agents" easier. and trusting agents is pretty important to adopting them. if you want to delegate something important to an agent or a human, you want some level of predictability; and determinism is just the extreme case of predictability.
-
Open-Sourcing Jailbreaking Defense Benchmarking Tool
By
–
We’re open-sourcing our code so others can build on our work. We hope it assists in benchmarking misuse risk, and designing defenses to safeguard against strong adaptive attacks. Paper: https://
arxiv.org/abs/2412.03556
GitHub: https://
github.com/jplhughes/bon-
jailbreaking
…
Website: https://
jplhughes.github.io/bon-jailbreaki
ng
… -
Best-of-N Jailbreak Technique Achieves Faster ASR on Claude Models
By
–
Best-of-N can be composed with other jailbreak techniques for even more effective attacks with improved sample efficiency. Combined with many-shot jailbreaking on text inputs, Best-of-N achieves the same ASR 28x faster for Claude 3.5 Sonnet and 70x faster for Claude 3 Opus.
-
Frontier AI Labs Responsible Vulnerability Disclosure via Forum
By
–
Before sharing these results publicly, we disclosed the vulnerability to other frontier AI labs via the Frontier Model Forum (
@fmf_org
). Responsible disclosure of jailbreaks is essential, especially as AI models become more capable. -
Jailbreak Attack Success Rates Scale with Compute Power
By
–
Attack success rates scale predictably with sample size, following a power law. More samples lead to higher success rates; Best-of-N can harness more compute for tougher jailbreaks. This predictable scaling allows accurate forecasting of ASR when using more samples.