Last Week in AI ACM on Facial Recognition, National AI Cloud, and Positive DeepFakes
Jul 10, 2020
A compelling discussion unfolds around the ACM's call to halt facial recognition technology due to ethical dilemmas and public backlash. The need for better dataset accountability in AI is emphasized, with MIT's recent actions showcasing a shift towards responsible practices. Delving into AI data management, the potential creation of a National AI Research Cloud is explored for enhanced collaboration. The transformative use of deepfake technology is highlighted, especially in supporting anonymity for vulnerable individuals while stressing the importance of regulation to harness its positive applications.
AI Snips
Chapters
Transcript
Episode notes
ACM's Call for a Moratorium on Facial Recognition
- The Association for Computing Machinery (ACM) urges lawmakers to suspend facial recognition use.
- This call is for a temporary moratorium until accuracy standards and regulations are in place.
MIT Removes Biased Dataset
- MIT removed the "80 Million Tiny Images" dataset due to racist and misogynistic labels.
- This highlights the importance of careful dataset curation in AI.
Building Better Datasets
- Datasets should be carefully constructed and filtered to avoid biases.
- The AI community is increasingly recognizing and addressing these issues.