How to Steal an Election
summary
Demonstrates concrete ways data science can swing results: using segmentation to 2,000 voters (31 marginal UK seats with margins under ~2,000 votes), redrawing districts via gerrymandering, rejecting postal ballots, and caging through ID rules to suppress turnout. Applies behavioral tricks (compensation, broken‑windows, Pygmalion, obedience), exploits a post‑factual public (the “£350m” claim example), and runs A/B and uplift testing with microtargeted social ads where no regulator checks content, optimizing spend on persuadables rather than sure things or lost causes.
AI-written summary · about
chapters (32)
- 0:00Start
- 0:59So, what are we doing here?
- 3:56Macro / Micro Scale
- 5:09Macro Scale
- 5:20Segmentation / Clustering
- 7:18Gerrymandering
- 8:24Tossing
- 9:22Caging
- 13:27Bribery
- 14:36Blackmail
- 15:24Behavioural Psychology
- 15:52Demo: I want someone killed
- 17:30Psychological Tricks
- 17:41Compensation Effect
- 18:15The Power of Names
- 19:06The Broken Window Effect
- 20:12The Pygmalion Effect
- 21:05Pressure To Conform
- 21:38Obedience To Authority
- 23:06System Justification Theory
- 24:20Vote Status Quo: Inject Fear
- 26:37Vote Change: Better Off/More Stable
- 33:44Simple ideas are deemed true
- 36:02How did we get here?!
- 42:39Demo: Tomorrow's Weather
- 48:42A/B/C Testing
- 50:08Between Measures Testing
- 50:55Uplift Modelling
- 53:30The SNP
- 54:23Silence The Media
- 58:01Cancel Culture
- 59:30Provide Your Own News
related talks
44mHow to Fix an Indian Election
52mHacking Elections For Fun & Profit
45mFrom Ballots to Code: Exploring Election Security
33mElection Night Analysis: Art or Science
51mHacking Ten Million Useful Idiots: Online Propaganda as a Socio-Technical Security Project
33mMeet the people fighting surveillance capitalism
01h 08mAI and Elections Disinformation, Deepfakes, Dystopia
22mFacebook data and machine learning predicting elections?