“Algorithms are opinions embedded in code.”
Cathy O’Neil told me that in 2019, back when most people still assumed math was neutral and #BigData would make our decisions cleaner, fairer, and smarter.
She had good reason to doubt it. Harvard PhD in mathematics. MIT postdoc. Then a quant at the hedge fund D.E. Shaw in the middle of the credit crisis, where she watched math get weaponized from the inside. What she saw there became Weapons of Math Destruction.
Seven years on, those models have learned to talk. They read résumés, score borrowers, flag “high-risk” neighborhoods, and now they can write a confident explanation for every call they make. We traded blind faith in big data for blind faith in #AI. Did we fix anything, or did we just make the opinions more persuasive?
In our 90-minute conversation, we get into:
Manifest Destiny, American exceptionalism, and why we don’t count the dead the Facebook memo arguing that connecting people is “de facto good,” even when it costs lives what predictive policing software like PredPol is really predicting whether #DataScience needs its own Hippocratic oath why I believe Instagram may be the worst weapon of math destruction of them all.
Near the end, Cathy names the one thing the machines can never supply. Of everything we covered, that part has held up best.
Watch or listen to my full conversation with Cathy O’Neil on Weapons of Math Destruction here: [ https://snglrty.co/4eU14zE](https://snglrty.co/4eU14zE)
‘Weapons of Math Destruction’ is a must-read book for anyone who thinks that algorithms are by definition a fair and unbiased way to produce a given result.