Before I start, allow me to thank you all for your super interesting comments on last week’s post. I love hearing and learning from all of you!
And now for our regularly scheduled program…:)
For about a year I’ve come across a few stories about beauty contests being judged by robots, that is Artificial Intelligence. By measuring people’s (let’s face it, women’s) facial structures these robots can accurately determine beauty. And given the judges are robots versus actual humans, we assume that the judgements are devoid of cultural or ethnic biases. I’ve read these stories with mild interest. But this time, a recent story stuck with me…and not in a good way.
I’m sure such a contest raises a whole host of issues for many of you. First, I imagine many of you oppose such contests — humans or no humans as judges. How can one judge beauty anyway? Second, so much of physical beauty emanates from within. I’m not talking about inner beauty. That’s a whole other subject. I’m talking about the energy, the light, the passion that springs forth from someone making then either more or less beautiful. Can a robot really judge that?
But those issues aside, my biggest concern with this contest is what is conveys about technology as a whole. The results of these contests showed how actually, how terribly biased A.I. can be!
According to this article in NextWeb, the contest drew over 6,000 applications from across 100 countries. And despite the obvious diversity of beauty that the robots were exposed to, out of the 44 winners, only a few were Asian, one was black, and the rest were white. None had dark skin.
WTF!
Are the foolproof “complex algorithms” that biased? I’m afraid so. How can that be? Easy. A human being has designed them that way. And that’s pretty scary.
I’m not saying the engineers behind the algorithms intentionally developed them to be pro white. It’s just that bias is going to inevitable when the majority of the brilliant brains behind the development of technology are of a certain gender, ethnic background or culture. No matter how much we try to rid ourselves of our biases, it’s super, super hard to erase the ones we don’t even know we have.
If technology can be biased when it comes to beauty, could it also be biased when it comes to truly understanding the user and what s/he needs? Or can it be biased with regards to places on the map or particular destinations? You get the point.
So what do we do? It’s not a new news that we should push harder for diversity our work places. It makes for diversity of thought and ideas. But it goes deeper than that. Diversity ensures that the seemingly “judge-free,” non-biased technologies we create actually hold up to that expectation. When humans are led by their biases, we forgive them. Because, well, it’s “human” not to be perfect. But technology shouldn’t make mistakes, right? It can’t be biased. So if technology, or in this case A.I., declares someone fit or beautiful or smart, well then it must be so! And the result? People deem the seemingly unbiased robots as arbiters of truth.
I’m not the first person to call for more diversity. It’s a MUST. But I’m also calling on all of us to challenge our notions of how “smart” our technology really is. Believe me, I LOVE technology. I’m benefiting from it in all aspects of my life, and, most, importantly in my job. But let’s realize that behind all technology are human beings. This recognition should mostly strike a sense of admiration in human kind. After all it takes a buttload of brilliance to be creating the amazing technology we have, and will have in the coming years. But it should also open our eyes to that fact that technology isn’t fool proof, totally unbiased or “right” all the time.
Let’s remember, human beings are the smartest, most elegant and beautiful “machines” that exist. We are complex, emotional and gorgeous in so many diverse ways. We don’t need a technology-driven beauty contest to prove that.