Sculpture: The Flybrary, Burning Man 2019 by Christina Sporrong

Have you ever wondered how 2020 will be remembered in history?

Dean Ginsberg

--

This weekend I saw a caravan of Trump supports protesting the legitimacy of the recent election results. It was clear to me that although we lived through the same experience just weeks ago, our narratives of what happened were wildly different.

We were both steadfast in our opposing truths, and it made me wonder, which narrative would be remembered long after we were gone?

I studied history at Oberlin College and was obsessed with the narratives and people that shaped our collective identity. That same passion drove me to be a CMO where it is my job to connect with people on a grand scale.

However, the more I learned about history the more I realized that there is no single narrative that accurately depicts what happened.

What role does Truth (with a capital T) play when we are all living in our own reality?

When historians recognize the bias embedded in the interpretations of their colleagues from previous generations, they seek new primary documents that shape new narratives. There is a whole academic focus on this called Historiography.

When I learned about historiography, it made me question the whole concept of truth.

I became obsessed with a simple question: what is more important, what actually happened, or the story we are told about what happened?

If you’re reading this you may be disgusted, as I was. Obviously what actually happened is more important! But, think about a time when you tried to mediate an argument between two friends, colleagues, or family members…

You likely found that two people can observe the exact same event and walk away with two wildly different interpretations of what actually happened.

If both of these people are still alive you could try to piece together the truth by talking to each of them, but your conclusion would inevitably be riddled with your own bias. Now imagine searching for the truth by piecing together fragments of diaries that are hundreds of years old and heavily dominated by one side.

This is the job of a historian.

It is riddled with so much bias that many believe we can learn more about how people thought in one period by studying the way they interpreted history from a previous period.

Trippy right? Now think about the fact that those narratives of history shape everything that we form our collective identity around, in the same way, our memories of childhood shape who we are as individuals today.

Which leads me to today…

We now live in a world where 2.4 billion people upload four petabytes (4 billion gigabytes) of data to Facebook EVERY DAY.

And that’s just Facebook, over 300 hours of video are uploaded to YouTube every minute, and 2 trillion searches are made on Google every year. Imagine if historians had that amount of data when piecing together what life was like in Colonial America or Ancient Rome!

I believe that in the future, historians will not ruffle through archives to find a single letter or diary. They will mine mountains upon mountains of data from almost every person on the planet to piece together our past.

Now the crazy part. No human can possibly sift through that much data. We will have to rely on artificial intelligence to shape our historical narratives.

The historians of the future will not be book worms or archeologists, they will be data scientists and AI architects.

This reality is both incredibly exciting and absolutely frightening.

On the one hand, it could lead us closer to historical truth, something we are far away from today. On the other hand, we are trusting a computer to define the narratives that shape our collective identity.

Despite my love of the Terminator movies and perpetual fear of Sky Net, I don’t think the actual computers are what we need to cautious of (at least at first.) What I find most concerning is the power this paradigm shift will put into the hands of the people who write the code behind these algorithms.

Just like the historians of today, the historical AI architects of the future will have their own biases. These biases will not be revealed by the primary sources they focus on, it will be embedded in the code that they write and will be far more difficult to discern.

Given that currently, only 12% of AI researchers are women, a small fraction are people of color, and all AI researchers exist in the tech bubble, we can expect an even more powerful wave of single-perspective dominance in the shaping of our history and thus our collective identity.

Although we’re talking about the study of history in the future, what we are actually talking about is how we will be remembered today.

The key to this problem is the question of who has access to the data.

Currently, historians rely heavily on letters of correspondence and diaries which are first owned personally and make their way into the public records over time.

What happens when that personal correspondence is owned by a handful of corporations?

Our ability to discern bias in these narratives will be dependent on our access to the data and the code behind the algorithms, both of which are proprietary and currently shielded from public view.

I’m sure there are smarter people than me thinking about this. I’d love to see what AI-driven history looks like when it’s led by a representative group of historians who have public access to data. If you’ve heard of anything or have thoughts of your own please share in the comments!

--

--

Dean Ginsberg

I tweet about products, startups, and my pleasantly plump bulldog Butch