This summer, I had the great privilege of attending EyeO (June 3–8 2018). Innumerable topics that encompass the intersection of Art, Technology, and Data were covered, but one common thread has left an imprint on my brain. That is: the Sisyphean 21st century task of disambiguating fact from fiction. That’s right…
The following lecturers touched upon this topic and forever changed my understanding of the 2018 landscape of fact, fiction, and storytelling’s role in deciphering one from the other.
PART 1: NUMBERS ARE MALLEABLE
On the first day, we discussed climate science at length. We (a very self aware room of liberal, number-crunching, data-visualization-making, coastal-living, self-ascribed nerds) attempted to break down the problems with human psychology. We looked at the facts, stats, charts, and graphs; then investigated the human power of denial, dissonance, disincentivization, and the hurdles of behavioral change. After 6 hours of discussion, ideation, and reflection, feeling a bit helpless, we ended with questions that I kept with me throughout the next 3 days of lectures:
Why don’t people believe statistics?
Are stories more powerful than numbers?
Why is denial more powerful than behavioral change?
Why do lies travel faster than truth?
…And what should we do about this?
The next day, Amanda Cox enlightened us with her talk These Lines Are The Same. She showed us that data, even in simple bar graphs, can be misread depending on the viewer’s own bias. She bravely revealed to us that in her department The Upshot at The New York Times they struggle with how to best represent datasets objectively. They experiment in meaningful and educational ways. In one example she showed data from the US unemployment report. The article allows readers to look at the chart with ‘Democratic Goggles’ and ‘Republican Goggles.’
The numbers are the same, but they can easily be bent to the will of anyone with an agenda.
Then she humorously showed us our flaws in clinging to round numbers. She drove the point home with a series of charts, one here showing the likelihood that someone in the ER gets checked for a heart attack, according to their age. As Amanda points out, “nothing radical changes from the age of 39-and-three-quarters and 40, yet here is the data:
Blacki Migliozzi, another NYTimes employee deeply rooted in citizen science, showed us his attempts to make climate science data even more interactive. His job is in part to ideate new ways of explaining the facts of climate science — mostly through longform data journalism. This reliance of quantitative results and ‘geeking out over the numbers’ from The New York Times admittedly raises questions about it’s ability to change minds. Within the backend underbelly these gorgeous data visualization lurks a reality that numbers are malleable, manipulatable, and therefore disputable.
Numbers are malleable, manipulatable, and therefore disputable.
PART 2: NUMBERS ARE USED UNETHICALLY
Nathaniel Raymond — a human rights investigator from Harvard, showed us how numbers’ pitfalls are not just their malleability, but their ability to be used unethically.
He elucidated instances throughout history — from cuneiform to computer vision drone surveillance, where accurate data collection was used for nefarious power-mongering destructive purposes. The most scarring example drew a line from the invention of the punchcard — originally created for census collection, to the numbers on Holocaust prisoners’ arms. There were many more examples; I encourage you to dig into their resources.
PART 3: DATASETS (WITH BIASES) ARE USED FOR MACHINE LEARNING
The third ‘angle of understanding’ (pun intended!) in the flaws of numbers, arises when you introduce Machine Learning. Meredith Whittaker (founder of AI Now and open research lead at Google) walked us through a series of examples showing human bias in machine learning corpuses. For all of us CS-loving, ML-summer-prancing, VC-money-grabbing humans that like to claim that anything that comes from a computer is fact, think again. The things we currently take as fact (datasets) are deeply rooted in human error. She highlighted example after example that are so rich they deserve their own article. It turns out that if we ‘move fast and break things’ at scale, then we have a broken system.
If we ‘move fast and break things’ then we have a broken system.
If we’re practicing Machine Learning, and especially if we are building companies and products off of it, we must take a moment and reflect upon these questions before we ‘hit commit’:
What data set did I train this on?
Who made that dataset?
Was the dataset itself biased in some way that will bias my results?*
*spoiler, the answer is probabilistically ‘yes.’
This 3-part view of how numbers can lie, be used unethically, and be controlled by our own bias is admittedly grim. So, should we just give up? Abandon society and run for the woods? Remove ourselves from all social media and quit our tech jobs?
I propose NOT doing that. I propose that we continue to be creatures of habit and do what we do best — tell stories.
PART 1: FORGET NUMBERS. WE ARE STORYTELLERS.
If quantitative won’t work, let’s use qualitative: anecdotes, speculative futures, cinéma vérité, and empathy. These are powerful tools that many EyeO speakers (including all listed above) are using in really effective ways.
The powerful women who run hyphen-labs showed a poignant piece addressing the opioid epidemic in America. A poetic mixture of quantitative and qualitative, they created an infinity wall of faces — every face represents someone who overdosed on opioids. Loved ones and the currently afflicted can visit, feel the gravity of the sheer quantity and and the lost humanity of this epidemic. It’s powerful. It’s a marvel in 21st century manufacturing. And you can watch this video about it and be moved:
Chris Cheung works to create speculative futures, where he addresses issues of slave/master relationships between humans and technology. His AG Collar is a fully functioning human collar that has embedded location tracking, geotagged audio guidance, and a timer. Note, these are all components of the smartphone ‘master’ that is in your pocket or hand right now.
Many other poignant storytelling pieces were presented — it’s EyeO after all! I won’t delve into the importance and impact of all of them, but here are a few more you can dive into on your own:
PART 2: IF NOT SCIENCE, SCIENCE FICTION.
The final keynote speaker, Ariel Waldman, drove the point home, that in the absence of societal readiness for truth, let’s make sci-fi. Ariel’s mission is to make space exploration accessible to all; she showed us the importance of experimenting, collaborating, and telling stories, even in the face of adversity and oppressive societies.
She told us a story about Johannes Kepler’s vision of a moon landing in 1608. The 17th century was when Galileo was convicted of heresy for proposing that the Earth revolved around the sun instead of vice versa — the truth. At the same time, Kepler was conducting thought experiments about human travel to the moon. His thoughts were deeply rooted in the laws of physics, but in order to not get arrested, he disguised his science as science-fiction, in his book ‘Somnium’.
For those of us who believe that technology can still create a better world given all of this manipulation, these projects of speculative futures, powerful emotional storytelling, and science fiction, are all paving a path. Even if we have to disguise our facts and science-fiction, lets get our messages out in any way we can.
Even if we have to disguise our facts and science-fiction, lets get our messages out in any way we can.
If we have to create a VR game of all black women scientist in order for people to see that reality, lets celebrate that hyphen-labs is doing that. If we have to create a game called ‘swipe for fact or fiction’ in order for people to practice deciphering between real and fake news, let’s play it. And if we have to wear collars around our neck at an art show to realize that our phones own us, lets participate.
If we allow ourselves to get defeated by the futility of numbers, statistics, and human bias, then it truly is a ‘post-truth world’. But if we do what we do best, and tell stories that are rooted in science, fact, and optimism, then we can drive humanity to where we want to go. Let’s drive, draw, whistle-blow when we need to, and keep making.
Thanks for reading. Thanks for commenting.
Please 👏 👏 👏 CLAP if you enjoyed it — that’s how things get upvoted and spread!
THANK YOU EYEO!