Ethical Artificial Intelligence to Advance Welcoming Communities and Identify Potential Communica...
summary
A pipeline called Auger data pipeline mines GitHub commits, issues, PRs and messages and runs discourse clustering, message novelty scoring and a pull request merge predictor; models retrain about every 30 days. Trained on 3,500 corporatized and 2,300 scientific projects, it flags a small set (dozens) from hundreds of thousands of messages for human review, finds higher negative valence in scientific repos, and groups projects into clusters that correlate communication patterns with project stability — useful for triage but still requiring human validation.
AI-written summary · about
chapters (7)
- 0:00Start
- 3:27How the Human Brain Works and How Machine Learning Works
- 8:36Discourse Analysis
- 10:01Pull Request Analysis Worker
- 11:48Toxic Messages
- 12:08The Signs of Healthy Communication as Opposed to the Signs of Unhealthy Communication
- 17:21Ethics of Machine Learning Algorithms
related talks
41mUsing AI to Build Fair and Equitable Workplaces
07mInterview
46mData Discussion - AI for Good
35mThe Podcast Pipeline Nobody Sees: Behind the Scenes with AI Agents and AI gen
45mGit push get an AI API
02h 03mBuilding on AI - Delhi meet-up
16mSelf Driving Products: Product Signals to Pull Requests
48mAI Impact on Cybersecurity and Privacy