New York recently passed a bill providing guidelines for the operation of automated hiring software in the city.
Allegedly, city councils aim with this legislation was to protect New Yorkers from biased AI.
But, from where were sitting, it looks like its going to do the exact opposite.
But the reality is that theres no magic way to remove bias from an AI system.
Algorithms arent magical spells.
Theyre usually math-based guidelines.
And, because computers and algorithms are designed and programmed by humans, they containinherent bias.
40% off TNW Conference!
All they can do is make guesses.
Some guesses are harmless, like when Netflix guesses youll like a certain movie and you dont.
you’re able to just watch something else.
Other guesses are bad, like when an algorithm decides youre not worth hiring.
AI absolutely cannot tell if youre lying.
And it certainlycant measure human intelligence.
Good idea:passing a bill that protects New Yorkers from predatory snake oil companies.
Bad idea:letting the softwaresvendorsdecide whether an algorithm is biased or not.
Violations of the provisions of the bill would be subject to a civil penalty.
But, those violations?
Its up to thevendorto conduct and report audits demonstrating their algorithms arent biased.
Quick take:This is like lettingThanosconduct audits and report compliance with New York City ethics.
According to the big purple dude, his use of the Infinity Gauntlet is perfectly ethical.
He did an audit and everything.
This doesnt protect citizens, it protects scammy AI companies.
HireVue gave a statement to the AP that sums it up best.
HireVuecurrently uses AIto allegedly assess candidates e-motions.
The company actually claims computer vision algorithms can determine a users emotional intelligence.
This, of course, isnot possiblebecauseAI doesnt have magic powers.
