Design with the End-user in Mind
SAFETY
-
Robust Artificial Intelligence: Four Steps Towards AGI
By
–
"absolutely brilliant" —Nobel Laureate Danny Kahneman The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence @GaryMarcus : https://
arxiv.org/abs/2002.06177 #AGI #AIDebate #AGIDebate -
ImageNetX: Identifying Vision System Failures at Scale
By
–
Even today’s best #deeplearning vision systems can fail when pose/lighting/background vary. Our work on ImageNetX is one of the first large scale efforts to pinpoint mistake types of in AI computer vision systems. Explore the dataset
-
Image Generation Quality Issues and Current Technology Limitations
By
–
Use better input pics, and even then many of pics are unusable and have artefacts. Reality of new tech
-
Large language models quantifiably tell us what we want to hear.
By
–
Large language models are, quantifiably, telling us what we want to hear.
-
Biased AI Influences Emergency Healthcare Decision Making
By
–
When making emergency health care decisions, people are “highly influenced" by biased AI recommendations: https://
bit.ly/3v49Y8v -
Trustworthy ML Workshop at ICLR 2023: Statistical Computational Limitations
By
–
A super cool and timely workshop to be co-hosted at #ICLR2023! When can statistical and computational limitations arise in the context of trustworth ML? Deadline is February 8, check out their unique two track system. https://
sites.google.com/view/trustml-u
nlimited/call-for-papers
… Look forward to being there! -
IBM Emphasizes the Importance of Responsible AI Governance
By
–
In the wake of increasingly accessible and sophisticated #AI models adopting the core principles of AI governance is now critical for the deployment of compliant, reliable, responsible AI. – IBM GM @dineshknirmal → https://ibm.co/3BOGgYS #ChatGPT #datascience #automation
-
Machine typing ChatGPT outputs bypasses verification systems
By
–
But what if we make a machine that types on a verified keyboard whatever ChatGPT says? It's hard to avoid that
-
Anthropic Plans to Share Safety Experiment Data
By
–
As we did with our 'red teaming' project (https://github.com/anthropics/hh-rlhf…), we plan to release the data from this experiment in the future to empower a broader set of people to build safer systems.