Consequences of an Insightful Algorithm
summary
Shows how deep learning and rule‑based code actually behave: a predictive algorithm detecting second trimester that exposed a teen, Shutterfly emails that reopen miscarriage grief, Facebook Year in Review resurfacing a dead child, Fitbit exposing sexual activity, Uber’s God View letting employees stalk riders, OkCupid and AdWords reinforcing bias (Harvard finds black‑identifying names 25% more likely to trigger arrest ads), Google Photos mislabeling people, a Mario net that learns in 24 hours — concrete failures, tools used, numbers, and how fixes did or didn’t work.
AI-written summary · about
chapters (11)
- 0:00Start
- 1:51ALGORITHMS FOR FAST, TRAINABLE, ARTIFICIAL NEURAL NETWORKS
- 2:24DEEP LEARNING IS A PARTICULAR APPROACH TO BUILDING & TRAINING ARTIFICIAL NEURAL NETWORKS. THINK OF THEM AS DECISION- MAKING BLACK BOXES.
- 3:23Behavioral Prediction Image Classification Face Recognition Sentiment Analysis
- 4:03Deep Learning relies on an ANN's automated discovery of patterns within its training dataset.
- 11:28INCREASE AWARENESS OF AND CONSIDERATION FOR THE FAILURE MODES THE EDGE CASES THE WORST-CASE SCENARIOS
- 11:48BE HUMBLE WE CANNOT INTUIT INNER STATE EMOTIONS PRIVATE SUBJECTIVITY
- 23:55Many regulations are in place to correct for systemic, structural bias against minority groups. Ask yourself if you want that "disrupted."
- 24:59UNDERLYING ASSUMPTIONS INFLUENCE OUTCOMES AND CONSEQUENCES
- 27:40WE MUST HAVE DECISION-MAKING AUTHORITY IN THE HANDS OF HIGHLY DIVERSE TEAMS
- 31:07PROFESSIONALS APPLY EXPERTISE & JUDGEMENT ABOUT HOW TO SOLVE PROBLEMS
related talks
26mConsequences of an Insightful Algorithm
30mInside AI: What Every Professional Needs to Know to Stay Relevant, Protected, and Ahead
24mDesigning for Transparency in Machine Learning
52mSome pitfalls in AI
12mHow we infuse our biases in the products...
40mTop Seven AI Breaches: Learning to Protect Against Unsafe Learning
04mYour data is being manipulated.
37mThe Shadows That Follow AI Generative Models