Artificial emotional intelligence is near

Artificial emotional intelligence is near PsychologyThere’s a simple reason why you need to understand how emotions work, and it’s this: emotions drive people’s decisions, not rational thinking. So if you want to understand how to build more persuasive tech, you’d better invest some time to learn how emotions work.

Emotions are so hard to study, many scientists ignore them

People are often surprised when I tell them that scientists don’t fully understand how emotions shape behavior. They’re even more surprised when I tell them that many behavioral scientists don’t even bother trying to understand emotion because they’re difficult to measure, and it’s much easier to study the external factors that shape behavior.

For instance, it’s easy to measure how technique X impacts on behavior Z. But it’s, extremely difficult to measure how technique X, impacts emotion Y, which shapes behavior Z. For this reason, most persuasion science focuses on external techniques, rather than internal emotional drivers.

Academics are divided on the science of emotion

Not only is it hard to measure emotions, but scientists don’t fully understand how they work, and worse, they’re divided into two major academic camps, each with radically different views on what they are, and how to describe them.

The first camp, which I’ll call the basic emotions camp, is led by academics who argue there are 6-14 basic emotions, and each emotion has its own neural circuit. They believe that events trigger emotions by firing-up the neural circuits tied to specific emotions. For instance, when your neural circuit for happiness is triggered, you feel happy. Or if your disgust circuit fires up, your body gets ready to deal with something gross, and you react. This approach originated with Charles Darwin, and subsequently, it has been advanced by numerous scholars, including Paul Ekman who is perhaps the most famous for advancing this line of thought.

The second camp takes a dimensional and neurological approach to emotion, with scholars arguing that emotions are far more complex, and can’t be classified into simple categories. Rather, emotions result from a complex interplay of neural circuitry, chemicals, and interactions that are shaped by a person’s life.

This second perspective is starting to take off more with neuroscientists and is most compatible with the psychological technologies and algorithms that we’re developing.

Recently, I picked up a copy of Dr. Barrett’s book, How Emotions are Made, thanks to a tweet from Roger Dooley (@rogerdooley). Though I just started, I wanted to share my early impressions, as a nice summary of the latest neuroscientific views on emotions, based on the argument that emotions are constructed in the moment by core systems that interact across the whole brain, aided by a person’s past experiences.

Good science leads to effective products

So, if you’re working with intelligent technologies, algorithms or strategies, you may be wondering, which scientific perspectives you should use? This is an important question to ask, because if you build on outdated scientific models, then chances are, you’ll be deploying second rate products.

In my experience, the best science leads to the highest impact products. But when it comes to emotions, there’s a practical problem here, because it’s extremely difficult to measure emotions, as the technology used to quantify them is not that reliable, hard to interpret, and you won’t always know what you’re measuring.

This doesn’t mean you won’t be successful. For instance, companies regularly make a killing by selling products based on flawed psychological models, such as Myers-Briggs, which doesn’t have much validation behind it.

Or in my digital psychology field, I see people selling psychology models that have zero scientific validation behind them. The bizarre irony, is that the best scientifically validated behavior change models are largely unknown, while some of the pop-psychology models with no validation get talked about most.

Through aggressive marketing, any second-rate product, service or model can become successful. However, if you care about building products that work, then I recommend building on the best available evidence.

More money is spent building technology based on the wrong model

With it comes to evaluations of programs based on the basic emotions camp, to the best of my knowledge, there are no amazing success stories.

Perhaps the most famous unsuccessful application of the basic emotion approach is the TSA’s POST program, which spent nearly one billion screening passengers for stress, fear, and deception (https://assets.documentcloud.org/documents/1697887/spot-referral.pdf). Overall, the program has been criticized by numerous evaluations as largely ineffective, despite its massive price tag.

Given that numerous technologies are grounded in the basic emotion camp, using the same beliefs behind the POST program, it’s no surprise that we haven’t yet witnessed the game-changing innovations, that we should expect, in this space.

When artificial emotional intelligence hits

I do, believe the day is near, when mind-reading, emotionally intelligent software, will read you like a book, and play you like a fiddle.

However, the breakthroughs will come from neuroscience, and some of the recent perspectives on how emotions work, based on insights from mammal studies, that are now being applied to humans.

When this day comes, we’ll enter a new era of artificial emotional intelligence, where the average person has access to a wide range of emotion reading and relationship enhancing support technologies; when we’re scrutinized by minority report type algorithms, subject to ‘nanny state’ pre-emptive healthy-development interventions, and living in an ecosystem where advertisements are hyper-personalized to the point where we’re more likely to feel annoyed when marketers haven’t even bothered to data mine our digital footprint, and infer the essence of our innermost needs.

I’ve spent almost three years working towards this future, collaborating with some awesome people to build emotionally intelligent algorithms, https://www.emotionalintelligence.ai, running emotional design workshops across North America (https://www.alterspark.com/e2), and now we’re adapting our tools to organizational development applications.

Why I’ve taken sides, with the dimensional and neurological approaches

During these three years, I largely rejected the basic emotion research, in favor of the dimensional and neurological approach. Why? Well, it fits our data across numerous studies, produces better predictive models, and it’s easier to apply, so there’s a good practical argument too.

Many of the emotionally intelligent technologies sold today, are based on the basic emotion approach, which many people don’t know, was inspired by Darwin’s original writings on emotion, which date back to 1872. This was groundbreaking over 100 years ago, and has been radically advanced by the basic emotion camp, but it’s time for our thinking to evolve.

Darwin gave us the theory of evolution, but it was Geneticists who showed us the details, how DNA drives the process. In the same way, Darwin also gave us our basic understanding of emotions, but now we need to turn to the neuroscientists to help us understand the details.

There’s a scientific revolution underway, that’s transforming our understanding of what emotion are and how they work. It’s only a matter of years before artificial emotional intelligence hits, and transforms our social order.

By | 2018-02-21T22:39:40+00:00 4 May 2017|Categories: Behavioral science, Persuasive design, Digital marketing, Research|Comments Off on Artificial emotional intelligence is near

About the Author:

Download your FREE e-book

Here's your e-book