Andy Ellis (CSO, Akamai) - Humans and Risk
•Als PPTX, PDF herunterladen•
0 gefällt mir•748 views
Humans are Horrible at Risk Management. Humans are Awesome at Risk Management. Humans are horrible at risk management! Have you seen the news about Florida Man? How are we even still around? And yet, we are still around. In fact, humans are awesome at risk management; we’re now the dominant species on the planet. Why? How? Andy will share his thoughts on why humanity has significant advantages in making rapid, generally correct risk choices. You will learn how risk choices that appear unreasonable from the outside may not be; to identify the hidden factors in someone’s risk choice that most influence it; find out how to help guide people to risk choices that you find more favorable.
Empfohlen
Empfohlen
Weitere ähnliche Inhalte
Ähnlich wie Andy Ellis (CSO, Akamai) - Humans and Risk
Ähnlich wie Andy Ellis (CSO, Akamai) - Humans and Risk (20)
Mehr von Business of Software Conference
Mehr von Business of Software Conference (20)
Kürzlich hochgeladen
Kürzlich hochgeladen (20)
Andy Ellis (CSO, Akamai) - Humans and Risk
- 1. @CSOAndy Humans and Risk Andy Ellis @csoandy www.csoandy.com
- 2. @CSOAndy Why do people make incomprehensible decisions? SecurityProduct Owner Anyone can be a villain in someone’s story. Modal Bias!
- 3. @CSOAndy A business conversation? Here is my project. Is it safe? Here is my dangerous plan! Sign off on it! Here’s an ISO checklist! Fill it out! Justify infosec budget with this makework!
- 4. @CSOAndy A business conversation? What do these phrases mean? I can’t be bothered to understand this. We’ll fill it out for you. We’ll do the work, just give us more budget!
- 5. @CSOAndy A business conversation? Is it done yet? All I care about is my schedule. Why didn’t you use ChaCha20- Poly1305? I’m smarter than you are, and I know big words.
- 6. @CSOAndy A business conversation? Is that a showstopper? I’m going to do this no matter what…. No, but we don’t recommend launch. I’m just CYA over here. You’re fine, but I won’t admit it.
- 7. @CSOAndy Humans are at risk management. badatrociousawfulconfusingincompetentincomprehensiblehorribleperplexing____________ @CSOAndy
- 8. @CSOAndy Humans are at risk management. awesomeawful @CSOAndy
- 9. @CSOAndy Observe Orient Decide Act Attention Processing Executive Function Coordination O O D A #KeyTakeAway Decision Making: The OODA Loop @CSOAndy
- 11. @CSOAndy CONSTRAINTS ATTENTION Proximity Novelty Urgency MODELS Context Framing Expectations RISKS Costs Fears Expected outcomes TRAINED RESPONSES Practice Repetitive Low risk HAZARDS Distributed social networks Fast information flow Virtual proximity Confirmation bubbles Hindsight expectations Obscure costs Complex returns Repurposed responses Dunning-Kruger O O D A The Power of Models
- 12. @CSOAndy CONSTRAINTS ATTENTION Proximity Novelty Urgency MODELS Context Framing Expectations RISKS Costs Fears Expected outcomes TRAINED RESPONSES Practice Repetitive Low risk HAZARDS Distributed social networks Fast information flow Virtual proximity Confirmation bubbles Hindsight expectations Obscure costs Complex returns Repurposed responses Dunning-Kruger O O D A The Power of Models
- 13. @CSOAndy Historical paranoia “Monkey on rope ladder” by Rachel Coleman Finch is licensed under Creative Commons 2.0 Generic | @CSOAndy
- 14. @CSOAndy Prisoner’s Dilemma Actual Prisoner’s Dilemma Different communities have different expectations! If we believe our “partner” will cheat on us, we’ll cheat first. Loyal Betray -3 -1 -3 -10-10 -5 -1 -5 30% 19% 13% 40% Loyal Betray Loyal Betray
- 15. @CSOAndy CONSTRAINTS ATTENTION Proximity Novelty Urgency MODELS Context Framing Expectations RISKS Costs Fears Expected outcomes TRAINED RESPONSES Practice Repetitive Low risk HAZARDS Distributed social networks Fast information flow Virtual proximity Confirmation bubbles Hindsight expectations Obscure costs Complex returns Repurposed responses Dunning-Kruger O O D A Risky Business
- 16. @CSOAndy Hazards used to be simple… @CSOAndy “Attention Walmart Shoppers” by Robert Couse-Baker is licensed under Creative Commons 2.0 Generic
- 17. @CSOAndy So we think risk calculations ought to be easy Loss $5M $5B Probability 10% / year .01% / year ALE $500K $500K Price of buying $50K $50K Maintenance $14K $14K Reduction in events 10% 10% Cost $26K / year $26K / year Risk Reduction $50K / year $50K / year Savings $24K / year $24K / year
- 18. @CSOAndy Qualitative Risk Matrix High Damage Moderate Damage Low Damage Very Likely Priority 1 Priority 2 Priority 5 Likely Priority 3 Priority 4 Priority 7 Unlikely Priority 6 Priority 8 Priority 9
- 19. @CSOAndy High Damage Moderate Damage Low Damage Very Likely Priority 1 Priority 5 Likely Unlikely Priority 6 Priority 9 Qualitative Risk Matrix
- 20. @CSOAndy … but now risks are more complex. Image Source: Jeep® @CSOAndy
- 21. @CSOAndy Cost Context Matters You are given one opportunity to play a game. A fair, 20-sided die will be rolled. You bet X; if your number is rolled, you keep your bet, and get back 20X; otherwise, you lose your bet. Your expected payout is thus 1.05. Would you bet $1? Would you bet $10? Would you bet $100? Would you bet $1,000? Would you bet $10,000? Would you bet $100,000? Would you bet $1,000,000?
- 22. @CSOAndy “You don’t know what you’ve got ‘til it’s gone.” @CSOAndy You value something by what you give up to get it.
- 23. @CSOAndy Peltzman Effect Coffey, Seamus. “The Peltzman Effect.” Microeconomics and Behaviour, October 2010, microeconomicsandbehaviour.blogspot.com/2010/10/peltzman-effect.html RISK REDUCTION PERCEIVED RISK
- 24. @CSOAndy CONSTRAINTS ATTENTION Proximity Novelty Urgency MODELS Context Framing Expectations RISKS Costs Fears Expected outcomes TRAINED RESPONSES Practice Repetitive Low risk HAZARDS Distributed social networks Fast information flow Virtual proximity Confirmation bubbles Hindsight expectations Obscure costs Complex returns Repurposed responses Dunning-Kruger O O D A The Spotlight
- 25. @CSOAndy
- 26. @CSOAndy@CSOAndy Image Source: Simons, D. J., & Chabris, C. F. (1999). Gorillas in our midst: Sustained inattentional blindness for dynamic events. Perception, 28, 1059-1074
- 27. @CSOAndy@CSOAndy
- 28. @CSOAndy
- 29. @CSOAndy CONSTRAINTS ATTENTION Proximity Novelty Urgency MODELS Context Framing Expectations RISKS Costs Fears Expected outcomes TRAINED RESPONSES Practice Repetitive Low risk HAZARDS Distributed social networks Fast information flow Virtual proximity Confirmation bubbles Hindsight expectations Obscure costs Complex returns Repurposed responses Dunning-Kruger effect O O D A The Response Playbook
- 30. @CSOAndy System 1 vs System 2 LEFT RIGHT LEFT RIGHT LEFT RIGHT LEFT RIGHT
- 31. @CSOAndy System 1 vs System 2 LEFT LEFT LEFT RIGHT RIGHT LEFT RIGHT RIGHT
- 32. @CSOAndy CONSTRAINTS ATTENTION Proximity Novelty Urgency MODELS Context Framing Expectations RISKS Costs Fears Expected outcomes TRAINED RESPONSES Practice Repetitive Low risk HAZARDS Distributed social networks Fast information flow Virtual proximity Confirmation bubbles Hindsight expectations Obscure costs Complex returns Repurposed responses Dunning-Kruger effect O O D A
- 33. @CSOAndy Humans are awesome at risk management. situationally @CSOAndy
- 34. @CSOAndy
- 35. @CSOAndy CONSTRAINTS ATTENTION Proximity Novelty Urgency MODELS Context Framing Expectations RISKS Costs Fears Expected outcomes TRAINED RESPONSES Practice Repetitive Low risk HAZARDS Distributed social networks Fast information flow Virtual proximity Confirmation bubbles Hindsight expectations Obscure costs Complex returns Repurposed responses Dunning-Kruger effect The End of the Right Situation? #KeyTakeAway O O D A
- 36. @CSOAndy Improvement through Adversarial Learning A D O O O O D A
- 37. @CSOAndy A D O OInjection Stealth Misdirection Maneuver O O D A OODA Attacks Mistakes Paralysis Distraction Blindness Complacency Confusion Ineptitude Surprising Misleading
- 38. @CSOAndy Improvement through Introspection A D O O O O D A
- 39. @CSOAndy OODA Improvements O O D A
- 40. @CSOAndy OODA Improvements O O D A Noise reduction Instrumentation Hazard review Model analysis Bayesian retrospection Planning Impact assessment Training
- 42. @CSOAndy Copyright Notification • Akamai Logo (Triple Wave Swoosh and Akamai Name) and Tagline are copyright and registered marks of Akamai. Slide template copyright Akamai. • All photos marked with copyright information, unless they are in the public domain. • All other content and assembly is licensed under the Creative Commons 2.0 Attribution License (CC-BY). Non-binding expression of intent: if you’re inspired by this, and just create your own, no attribution required. If you’re copying graphics, Stephanie Sullivan of Akamai is the originator. If you’re copying texts, Andy Ellis of Akamai is the originator. (Note: If you’re copying photos, their copyright holds).
Hinweis der Redaktion
- Akamai Logo (Triple Wave Swoosh and Akamai Name) and Tagline are copyright and registered marks of Akamai.
- Labels build out from just the previous not injection/stealth