And now #ChatGPT is banned in @NYCSchools due to “concerns about negative impacts on student learning, and concerns regarding the safety and accuracy of content.” https://
wsj.com/articles/chatg
pt-banned-in-new-york-city-public-schools-over-concerns-about-cheating-learning-development-11673024059
…
SAFETY
-
ChatGPT Banned in NYC Schools Over Learning and Safety Concerns
By
–
-
Thorough Mitigation for Prompt Injection in AI Systems
By
–
It’s premature to call prompt injection solved, but this is the most thorough mitigation I’ve seen yet.
-
Refund for AI Content Reactor: Tool Too Powerful for User
By
–
Just sent a guy a refund for AI Content Reactor. It was too powerful. Felt like cheating. Didn't sit right with him. He decided not to use the information. I'll take it as a win.
-
Robust Estimators and Exponential Mechanism for Parameter Scoring
By
–
The notation is a dense, but the algorithm is simple. The score of a parameter vector θ (wrt a dataset) is the minimum number of datapoints that must be changed to make the robust estimator output (approximately) θ. Feed this score into the exponential mechanism. 8/n
-
Robust Algorithms Enable Private Algorithm Design
By
–
In our work, the main conceptual result shows the other direction: a robust algorithm can be used to design a private algorithm. 7/n
-
Differential Privacy Frameworks for Robust Statistical Estimators
By
–
Most have been instance-specific, e.g., here's a particular robust estimator, let's privatize it. Some have proposed broader frameworks, e.g., https://
arxiv.org/abs/2111.06578 by Liu, Kong and @sewoong79
, or https://
arxiv.org/abs/2112.03548 by @praveshkkothari @pasin30055 @ameya_pa 5/n -
Private Algorithms Automatically Guarantee Robustness Properties
By
–
I like these works, but it would be nice to see something based solely on robustness or privacy, rather than properties of the algorithms themselves. Here's a work by @kris_georgiev1 & @Samuel_BKH that shows private algorithms are automatically robust https://
arxiv.org/abs/2211.00724 6/n -
Hand-coding fixes to machine-learned models ultimately fail
By
–
Hand-coding fixes to massive machine-learned models always fails in the end.
-
Generative AI Best Practices to Prevent Deepfake Crisis
By
–
Both these provisions are in line with global emerging best practices on how to continue using generative AI tools without creating a deepfake crisis.