Social Engineering Blogs

An Aggregator for Blogs About Social Engineering and Related Fields

Emotion in an Artificial Intelligence World

It is an almost omnipresent fear these days that technology is degrading human connections, but could we leverage that same technology to foster closer emotional ties?

We already have enough trouble reading each other’s emotions, and this becomes even harder when we communicate over long-distances, whether through email, phone calls, or even video chat. We are not face to face and cannot develop that emotional recognition, as technology, distance, and distracting stimuli add up to interfere with our emotional connections.

Instead, many people compensate with uniform online expressions, like emojis or acronyms conveying laughter or sadness, such as the classic LOL (Laughing Out Loud). While these can make progress in expressing emotions, they cannot replace actual, in-person laughter and lack a great deal of emotional nuance. Moreover, the problem seems intractable, given that modern society forces us to communicate over great distances in order to stay in touch.

That’s where developers like True Emoji, Affectiva, and Brain Power come in. These companies are using a concept called Emotion AI which attempts to give computers and other forms of artificial intelligence the ability to recognize and understand human emotions. It was Affectiva that developed a form of Emotion AI, and this software has been used in fascinating ways to teach computers to recognize facial expressions and understand emotions.

True Emoji has used this to address the problem of online communication. They have developed an app that reads a user’s facial expression and develops corresponding personalized emojis. This allows users to select an emoji that fits with their own emotion, rather than relying on a clichéd list of preset icons.

As True Emoji CEO Sumesh Dugar points out “The Internet has created a huge divide between emotions and communications… How many times have you shown surprise just by sending an emoticon?”

Similarly, Brain Power has put this principle to use trying to teach autistic children how to better recognize emotions. While autistic children have trouble recognizing emotion, this does not mean that they lack empathy. As we discussed in a previous blog, this just means that they can benefit from being taught emotional recognition skills!

Brain Power’s work focuses on the creation of interactive games that respond to the emotions of the participants. Joey Salisbury, the Brain Power director of software development, describes these as “augmented reality games with the goal of creating new, engaging ways for families to play together and grow closer while learning about emotions.”

These are just two applications of this burgeoning field. By abandoning an age-old divide between technology and emotion, we can allow the two to build on each other, creating emotionally intelligent machines and using technology to foster social ties.

For more information on this subject, see our previous blogs here and here.

Why Do We Smile?

You’ve probably heard that when you’re smiling, the whole world smiles with you. In reality, that’s not exactly the case.

According to researchers, Americans smile more than residents of other countries, and those other countries don’t seem eager to follow our grins. For some, persistent American smiling can be disconcerting.

To learn more, we talked to David Matsumoto, professor of psychology at San Francisco State University. He studies micro-expressions and non-verbal behavior.

LISTEN NOW

Ambiguity in Facial Expressions

How good are you at detecting emotions?

Followers of this blog have by now read extensively about basic emotions and the many characteristic features present in each. Perhaps this has made emotional recognition seem pretty simple. All we have to do is look at the lips, eyebrows, and other facial features, and we can conclude that someone is angry or sad, right?

Unfortunately, the reality is not that simple. Our brains do not deduce emotional states so rigorously (though they can be trained to!). Instead, we come to an immediate intuition as to another person’s expression from a broad interpretation of their overall facial features.

Moreover, we are often not very good at recognizing expressions. Untrained individuals generally have a difficult time identifying facial expressions, relying instead on feelings of empathy to come to understand other people.

This discussion is relevant, because not all expressions are even as obvious as our prototypes suggest. Certainly, there are universal basic emotions, such as fear and joy, but are all emotions as clear as being purely one or the other? Instead, many expressions, in reality, are ambiguous or offer subtle differences in intensity.

A recent study from the California Institute of Technology sought to explore these ambiguous expressions by analyzing the role that our brain’s amygdala has in making judgments about ambiguous or intense emotions.

These researchers analyzed brain activity within the amygdala when patients were shown pictures of people expressing fear or happiness, at different levels of intensity. They were also shown neutral or ambiguous emotions. Interestingly, two distinct groups of neurons responded to the facial expressions.

One of these neuron groups would activate intensely when exposed to strong emotions but was more muted during exposure to moderate or subtle expressions. Different neurons within this group correlated with fear and happiness. The other neuron group fired according to perceived ambiguity, regardless of the expression displayed.

The very fact that our amygdala has such an active role in identifying both intensity and ambiguity in emotional recognition helps better understand why emotional recognition can be so difficult. The amygdala is deeply connected with anxiety and fear centers, both of which infamously contribute to failures in recognizing emotions.

Study co-author Ueli Ruthishauser elaborated, saying “Researchers at multiple institutions are currently evaluating whether deep-brain stimulation of the amygdala is effective in treating severe cases of autism or post-traumatic stress disorder.”

As science continues to unravel the neurological underpinnings behind emotional recognition, we better understand how failures at reading these expressions can be solved and addressed.

In the meantime, check out this blog for more information on challenges to emotional recognition, or work on your own skills with Humintell’s training packages.