I’m reading the “What if” book from the XKCD creator.
What If?: Serious Scientific Answers to Absurd Hypothetical Questions is a collection of many of the blog’s most popular answers, along with brand new weird and wonderful questions.
I really like it. You can find the book online if you want to buy it and read it yourself. It made me think about some stuff. The biggest question that came to me, is the following:
How would the internet see humans if it was a little child and knew only how to look on Google for information?
Here is what I did to find out what the child will see when doing this:
- Go to Google Image through a Proxy and with an Incognito window
- Search for word and it’s opposite
- Take a screenshot
Why Google images? Because one picture is worth a thousand words. And all the children books are full of.. pictures! So, I opened up an incognito tab and browsed the internet through many proxies.
Looking at the results, I decided to write this post. Up to you to tell me what the child would learn from all of this. I wrote some comments, but you decide in the end..
Note: There are some uncool pictures, so if you’re more sensitive than a box of wet soap, don’t read further, close the tab, close the window, and close the computer.
People from different continents
I first wondered, how the child will see people from different continents; it is a first segmentation, geological one, of mankind that is visible from outside the globe. I tried the words “European”, “American”, “Australian”, “Asian” and “African” and here is what came up:
Result for European:
Result for American:
Result for Australian:
Result for African:
Result for Asian:
Result for Antarctican:
When seeing this, one thing was clear: there is a big difference between the European, American, Australian group and the African, Asian group. There is almost not a single human face on the first three, whereas there are a lot on the last ones. Why is that? It is a really curious phenomenon. I know it is only SEO but the underlying message is another one, more profound and cultural. It would be good to analyse the origin of each picture (website it is hosted on) in order to make some reasonable assumptions, and most of all, what would the child learn by googling this? I wonder.
People with different social status
Result for Poor:
Result for Rich:
This example is the most interesting one, it really nails it. Rich = Money, whatever the “Feel good videos or sayings” will try to convince you of. On the “poor” side we see groups of people, on the “rich” side, people alone, and money. What would the child understand?
People with different religions
I also wanted to see what’s about religions
Result for Christian:
Result for Muslim:
Result for Hindu:
Result for Jew:
Result for Buddhist:
I leave the interpretation up to you.
People with different gender
Result for Boy:
Result for Girl:
Ok, a boy is a little male human, a girl is a grown up female human with makeup and filters on her picture. Will the child think it is a boy or it is a girl, according to this?
Alone vs Together
People are alone and isolated. Stick figures are together.
Alive vs Dead
Being alive is a word like another, being dead is being a zombie with blood coming out of every possible pore.
Happy vs Sad
Symbols are happy, people are sad. Happiness is still not defined. I myself question the concept of happiness by seeing this.
Love vs Hate
Both are symbols, a bunch of letters, undefined, without feelings.
Old vs New
The old opposite of new.
Smart vs Dumb
I would not like someone to call me smart again.
Cold vs Hot
How would the child understand “it’s hot today”.
Good vs Evil
The difference between good and evil, is color.
Beautiful vs Ugly
Landscapes are beautiful, people are ugly.
War vs Peace
There are more people at war than people in peace.
I think these screenshots, are just a bunch of pictures, but they can tell us something about us. If they are in this order it is for a reason. If there is a sociologist out there reading this, or any qualified person, I would love to talk about it.
This is important because if ever we want to build an AI, and I do, we have to know how to teach it, and what to teach it. Language is deeply tied to emotions, feelings and context. I think solving this issue would be a big step forward. As we see, all sources of knowledge are biased. This was just an example of Google, but I bet the other sources behave the same way.
Besides, it’s fun to see that many stereotypes are confirmed when using this process of “learning”. We tend to put ideas and symbols before people and feelings, the latter we barely know how to define ourselves.
Some fun stuff to look for yourself: wet/dry, mature/immature. haha! Tell me your antonyms :)