Evil Digital Twins

They know us better than we know ourselves

“This group has a 72% conversion if shown Series 12B.”
“72? No. We need to sell 5,000 more by EOD to hit target, right? Ok, run the top twin categories again. Find 8,000, to be safe, and add a decent mood shift tolerance. I want at least 90% conversion from the new twin population before you send the actual blast.”

Does this sound like a sci-fi corporate sales scene? Or one just a little further into the future?

Current A/B/C+ testing — running multiple scenarios and messaging to see what causes someone to buy or use or watch something — is not far off.

  1. Swap in “there’s a probability of” for “there is”
  2. Drop the conversion estimate to less than 50%
  3. You’re at the present state of targeted advertising. You might hear a version of the above orders in any given Wall Street-style exchange in modern day — maybe at a sales conversion meeting at a modern day, hyper-aggressive corporation.

When enough of our thought process is mapped out, the next choice we make is predictable. Friends, partners, therapists, our own self-reflections — they all recognize our instincts and programming, and how we will likely respond in repeatable situations.

They understand us from getting to know us. Sometimes, they know us better than we know ourselves.
Photo by NeONBRAND

Artificial Intelligence systems can listen, observe, and then rapidly create a profile of our behaviors. So far, these systems have mostly looked for what kind of consumer or voter we are.

Replicas of our cognition are being assembled — not to impersonate us, but to predict and steer our decisions.

We are near a future where a particular kind of data science will be accurate enough to determine individual reactions — without involving the person. All it requires is their digital twin.

Welcome to the Era of Human-Digital Twins

#TBT ‘Mad Men’s’ Belle Jolie Lipstick Ad — shared on Yahoo Beauty 4.2.2015

Propaganda and marketing shift our perception of the world to fit the desired model. We’re convinced we should see the world a certain way. But consumers and citizens catch on. The Mad Men of the world were forced to get more subtle, less obvious, and tell more human stories.

Now, postmodern meta-marketing openly mocks the entire concept of advertising things to people. Being cool, self-deprecating, and “over it” sells more than function or fear.

Digital manipulation goes the other way. Subtle changes in the digital world are harder for us to pick up. Different, direct techniques can be used. Those “creepy” ads that seem to know what you just talked about — and certainly what you just searched for — are just the start. Like filtered banner ads, they’re already outdated and ignored.

The future of human manipulation lies in cloning digital twins.

Digital Twins are a “bridge between the physical and digital world.” They are used to run simulations, to test effects on widgets and factories and supply chains in similar environments to the real world. Except the tests are in a digital world, on clones of the physical thing.

Human-digital twins are similar. They are recreations of core, individual traits of a person. They act like us to accurately predict our reaction to new events.

No, they are not a complete mapping of a consciousness — neuroscience and philosophy still aren’t close to the core of that, much less AI. Human-Digital Twins only contain the elements of a person’s mind that interact with their creator’s platform.

And they have great potential to be used for evil. Our twins can be tested without us knowing. Futures can be predicted — and then found to be “creatable.” The new decision target creates a DeepFake of our minds.

Our digital clones will make our choices for (before) us.

Mapping minds

An effective manipulation system first maps out the perceptions of a population based on demonstrated behaviors. Then, it determines how an outcome goal could be produced in an individual model of perception. The changes needed to reach this goal become the manipulation campaign milestones.

Can this be applied at the individual level?

The human brain is hard to rewire. It’s why bad habits are so hard to break. It’s why addiction to technology can consume entire generations, whether or not they grew up as digital natives.

Data collection has already rendered baseline profiles for our digital consciousness. A few more learned conclusions on our consistent reactions and thought patterns, and our personal take on the universe is defined.

It’s not science fiction. All our current AI and computing systems need are some malevolent owners, some compliant engineers and a pipeline of fresh historical data.

And the Internet does not forget. We have social graphs and indexes of the entire web. We’re one measure of granularity away from individual graph databases, where the connected nodes represent the pathways of our thought patterns.

Human-Digital twins are born from this unsexy joining of bits and neuron clusters. They expand as we share more of our thought process online.

From “Brain mapping: New technique reveals how information is processed” by Erik Robinson of OHSU

The Psychology of [HD] Twins

Our identities and consciousness fall victim to cognitive capture. It keeps us in smaller definitions of individuality. “Thought leaders” signal what is ok to believe in and re-tell. Our unique combination of algorithms is forced into smaller, simpler boxes and categories.

We lose complexity so our evil digital twins can gain prediction accuracy for our next moves.

With no new ideas or preferences, there is no art, there is no true discovery or creative flow and sharing between changing human minds. The flood of images and short messages overwhelm and distract us. We do not notice the evil twin forming in the mirror.

“The [twin] process is one of the major instruments of social stability.” — the Director, 1st chapter, Aldous Huxley’s Brave New World
Photo by Tyler Mullins

In 1970, Alvin Toffler wrote in his book FutureShock of the “Modular Man.” We were all moving so fast, meeting and befriending and then leaving again — there was no time to fully understand each other. We could not share a deep view. We had to share select modules to stay light and agile and always moving into the future.

A manifesto of the future’s effect on us.

Fifty years later — how shallow and digital our modules are now. Social media thrusts us into the disparate fray of millions, even billions of voices. We whisper and yell and pretend to be happy and see beautiful things from a curated distance. We keep some form of connection with a number and spread of people thought unimaginable in 1970.

But we are still here.

We survived, are surviving, an ongoing Future Shock.

These tidepools of small talk and predefined preference — these are what our Human-Digital Twins study, and they don’t need to be any deeper. We become them when we repeatedly wade in the same, safe and shallow waters, afraid to explore the deep beyond the waves. Our twins stay with us onshore. They show us and tell us what they know we like, because they tested it on themselves. They are strange and evil.

Following the crowd is easy. Evil twins like easy — they were designed by the same models and manipulators that created the binaries of like and dislike, of showing half-assed support with a share or wordless emoji. If it’s too easy, you’re feeding your evil twin. They serve the man who made them, not you. They don’t keep [your] secrets. They will hold you on the shore, convince you it’s better there, follow you around like a clingy, bossy, shitty version of your online self.

Now, the Internet of Things’ version of digital twins are pretty boring — and extremely useful. When you clone hospitals for example, you see fatal bottlenecks in the system. They save lives. That’s not evil.

But we aren’t things. Human-digital twins are scary because we’re so human-centric in our view of the universe. We are the pinnacle of complex thoughts and individuality. We have freedom. We are “unknowable” things.

From the Richmond-Times Dispatch, Sophie Blystone’s 2014 take on dating a grown man living in his parents’ basement.

The HD Twin feels like a private investigator who has followed us around over the years. He’s started to predict our next moves — and he’s doing it by watching our data moves, presumably from some digital basement, eating old meatloaf and Cheetos and not paying rent.

Get a life, Evil Twin. Get your own life and stop watching mine.

So #GetOut, right?

#deletefacebook #quitsocialmedia #longlivetheluddites

Well, that’s not very practical. I won’t re-beat the dead horse of digital detoxing or going anti-tech. Technology is a toolbox; some tools are both made from us and kept from us, just out of our reach.

Photos by ev and 1AmFcS, respectively

Instead, consider what these Digital Twins are actually learning about us.

Daniel Kahneman’s 2011 book Thinking Fast and Slow still holds true for me. He describes two systems of thought in humans:

  • System 1 is the automatic, the reactive, the easy.
  • System 2 is the reflective, the articulate, and it takes effort.

Human-Digital Twins study our System 1. We can override and control our System 1 with our System 2. In this way, we have the final agency on our thoughts in System 2. What we choose to represent will be carried out by System 1. The actions turn into data points as we express ourselves across listening platforms. Our Digital Twin studies and inherits this System 1 character data, then repeats it.

Now, I do not think it's healthy to constantly “outsmart” and bluff our Twins in some kind of exhausting, endless digital chess match. Kahneman would say it's impossible to use System 2 this much. There’s a balance between protecting your identity and living your life.

Photo by Anh Le
Muting yourself is self-imposed censorship.
Consciously misrepresenting yourself is self-imposing the stress of a politician.

The Twin movement is not some conspiracy hatched by a few, dark corporate masterminds. It’s a byproduct of all the new data coming online about from sensors on things, from the up-and-to-the-right movement of bit-sharing among people. As the complexity of our twins increase, so does their utility for the new business models, ones built entirely on AI + Data.

Netflix may present a truly customized set of choices to us for consumption — a selection of films and shows only for us — even if the goal is to keep us consuming. A smaller niche company may use the same data to help us realize and confront mistakes we keep repeating.

In an odd way, Human-Digital Twins show the importance of understanding individuals. Labels and stereotypes could become less common if we map out and appreciate the complex mind for other, less evil purposes.

We can preserve our identity and avoid the chess game by leaning into the complexity of our System 2. Our Digital Twin does not have to be the backseat driver of our life.

Stay weird.

Stay weird and keep them guessing.