This morning, BuzzFeed News reported that some iPhone users have noticed Siri, Apple’s virtual assistant, gives an “odd and offensive response when asked ‘What is an Indian?’”
Siri would respond with a Wikipedia page extract for Indian, but say “They are a bit brown and they smell like curry and eat it.”
I don’t think I need to say that this has already prompted people to get upset. And this level of racial stereotyping is completely unacceptable, especially for a company that prides itself on a commitment to diversity. It probably doesn’t help that the tech industry is under scrutiny for diversity issues.
But apparently, that was enough for Siri to have stored an internal copy (known as caching) of the page content and start delivering offensive results.
How was it that Apple just happened to cache a version of the Wikipedia page that existed for only three minutes, when the versions both before and after the vandalism had existed for days?
Or was it a deliberate act?
Is it possible that someone has found a way to force Siri to update its caches of Wikipedia pages, and can now make Siri return offensive or even incorrect responses to users’ questions?
Because that would be seriously problematic: while being offensive is a serious thing, giving people wrong information to act upon would have even worse consequences.
What if Siri started telling people the wrong times or locations of their polling places on Voting Day?
You could systematically disenfranchise entire voting districts by lying to voters about when they could and couldn’t vote.
Apple needs to respond quickly to this incident. We need to know how this could have happened, and whether or not it’s possible to deliberately craft online content and manipulate Siri’s responses.
As we become more and more reliant on our technology products to function in society, it’s time we held these products to higher standards.
We would demand water companies explain how our water supply line was contaminated with pathogens and harmful chemicals, and hold them to account for it.
Why do we not do the same with information companies and our information supply lines when they are contaminated with useless, derogatory, or false information?