Speaking of position, does it matter if we are looking at the situation from an agentic perspective? For example, in this case we are addressing creators building and changing the situation itself. I'm not going to deploy a robot that actually feels pain, for example.
SAFETY
-
AI-Generated Content: Authenticity and Detection Challenges
By
–
It's not real, it was created using AI.
-
LLM Hallucinations: Origins, Mitigation Strategies, and Open Problems
By
–
Great tech talk on subtleties of LLM hallucinations by @johnschulman2 : where they come from, how to mitigate them, remaining open problems. https://
youtube.com/watch?v=hhiLw5
Q_UFg
… -
AI Safety Standards: Development Approach Discussion
By
–
I completely agree with this approach.
— Marek Rosa | European🇪🇺 | South African🇿🇦 (@marek_rosa) 27 avril 2023
Safety standards for AI systems are crucial and should be at the forefront of our discussions. It's important to ensure these systems are safe and reliable for all users.
However, pausing the development is not the right discussion.… https://t.co/ntMHK4J69XI completely agree with this approach. Safety standards for AI systems are crucial and should be at the forefront of our discussions. It's important to ensure these systems are safe and reliable for all users. However, pausing the development is not the right discussion.
-
Watermarking AI Text: Detectability and Open-Source Model Implications
By
–
Wouldn't be surprised if they soon do. But:
a) It seems likely that watermarked or detectable text can be (auto?) paraphrased into something that is no longer detectable
b) Watermarking wouldn't affect whatever future powerful open-source models that choose not to watermark -
Adversarial Robustness of VAEs at AISTATS 2023
By
–
Asif and @AmosStorkey at their best telling us about Adversarial
Robustness of VAEs. #AISTATS2023 -
We Are Obliged to Build AI We Are Not Obliged To
By
–
"We are obliged to build AI we are not obliged to." – Practical and rational talk by @j2bryson
-
U.S. Officials Raise Concerns About AI’s Cybersecurity Implications
By
–
AI’s Effects on Cybersecurity Concern U.S. Officials #cybersecurity #ai #rsa https://
wsj.com/articles/u-s-o
fficials-raise-concerns-about-ais-cybersecurity-implications-f32496ed
… -
Snapchat Launches ChatGPT-Powered AI: Safety Concerns for Kids
By
–
Snapchat Debuts ChatGPT-Powered Snap AI: But Is It Safe For Kids? https://
forbes.com/sites/bernardm
arr/2023/04/26/snapchat-debuts-chatgpt-powered-snap-ai–but-is-it-safe-for-kids/?sh=5227c30e68e2
… -
Congressional Meeting on AI Regulation and Public Safety
By
–
Met with @SenSchumer and many members of Congress about artificial intelligence regulation today. That which affects safety of the public has, over time, become regulated to ensure that companies do not cut corners. AI has great power to do good and evil. Better the former.