Studies show AI models can predict patient demographics including race from medical images. HAI faculty affiliate @james_y_zou discusses the real concerns — and possible benefit — in this: https://
stanford.io/3KuVSVZ
SAFETY
-
AI Models Predict Patient Demographics From Medical Images
By
–
-
AI Misidentification: When Algorithms Flag Innocent People as Threats
By
–
What would you do if AI flagged you as a terrorist? “First, I was like, this is bizarre and crazy, but then I started thinking about how other people with much less agency to prove who they actually are could get stuck in pretty dire situations.” https://
nyti.ms/3Qq2WqF -
AI Model Accurately Detects Typing by Listening to Keystrokes
By
–
Incredible: This AI model can determine what you are typing just by listening to your keystrokes – with 95% accuracy:
-
SAS VP discusses real AI risks and ethics concerns
By
–
SAS VP of Data Ethics, Reggie Townsend, talks to Fox5 New York about the where the real AI risks currently exist. http://
2.sas.com/6012PdLQi -
Generative AI blurs human-AI distinction and enables vulnerability
By
–
Thanks for the food for thought @SpirosMargaris
. Reminds me the movie "her". "Generative AI has made it harder for many people to distinguish between what came from AI and what came from a human. AI can also offer a space for people to be vulnerable because they can receive -
AI Worst-Case Scenario: Ethics and Digital Transformation Risks
By
–
#AI: The worst-case scenario https://
bit.ly/46ldAUY #ethics #DigitalTransformation -
LLaMA2 Responsible Usage Guide Released by Meta
By
–
LLaMA2 responsible usage guide(PDF): https://
scontent.fnbo1-1.fna.fbcdn.net/v/t39.8562-6/3
65271716_1020127512503290_6433760642443597145_n.pdf
… Guide website: -
The limitations of AI content detection and the importance of critical thinking
By
–
Je pense que à terme il faut pas compter là dessus, la seule soltion restera notre esprit critique. Même openai a supprimer sa solution de texte généré par ChatGPT parce que ca fonctionné pas.
-
Modern Turing Test: Measuring AI Capability Beyond Language
By
–
We need a Modern Turing Test that can test what an AI can DO, not just what it can say @rhysblakely @thetimes on the idea of Artificial Capable Intelligence and the Modern Turing Test.
-
AI Safety: The Room Confinement Paradox
By
–
The staying is safe. It's the leaving your room that's dicey. 🙂