29 May 2018
AI and emotional intelligence: blurring the line
Can artificial intelligence be emotionally intelligent?
For many, the answer to this question would be a resounding “no!”. However, like so many aspects of the rapidly-evolving field of AI, the line separating artificial intelligence and emotional intelligence might not be quite as clear as most would expect.
What is emotional intelligence?
Psychology Today defines emotional intelligence as: the ability to identify and manage one’s own emotions and the emotions of others. It is generally said to include three skills:
- emotional awareness
- the ability to harness emotions and apply them to tasks like thinking and problem solving
- the ability to manage emotions, which includes regulating your own emotions and cheering up or calming down other people.
We are all, every one of us, a beautiful and unique snowflake. Human beings and their emotions are complex.
So much so, in fact, that many human beings, with armies of counsellors and psychiatrists at our disposal, mountains of self-help books at our fingertips, and thousands of years of evolution at our backs struggle to understand our own emotions, let alone anyone else’s.
Knowing that, what hope could a machine possibly have of unravelling the mysteries of the human emotional spectrum and developing the 3 key skills of emotional intelligence of its own?
The emotional awareness equation
To a learning machine, different emotions and the human behaviors resulting from those emotions amount to little more than a series of data points in an equation waiting to be solved.
Boston-based startup Affectiva has recently been breaking ground on an Automotive AI system capable of recognizing seven emotional metrics and as many as 20 facial expression metrics in drivers and passengers with the goal of increasing driver/passenger safety in modern vehicles.
And Affectiva isn’t alone.
The quest to develop AI capable of recognizing and identifying human emotions based on facial expression, tone of voice, and body language is in full swing. It is widely regarded as a key step in the development of the human-AI interface, and many of the field’s best minds have been put to the task of helping artificial intelligence solve the emotional awareness equation in a practical way.
In his article The Rise of Emotionally Intelligent AI, Mikko Alassaarela posits that AI can learn emotional intelligence through much the same process we humans do.
“Our emotions and feelings are organic algorithms that respond to our environment. Algorithms, that are shaped by our cultural history, upbringing and life experiences. And they can be reverse engineered.”
The process by which human beings learn how other human beings respond to external stimuli is not so different from the process by which machines can do the same. Like toddlers in their first day of kindergarten interacting with a strange group of children for the first time, it’s simply a matter of observing the responses of others and committing those responses to memory.
Indeed, the processes that allow we humans to develop awareness/understanding of the myriad emotions in ourselves and those we interact with is hardwired and instinctual, but such skills appear to be well within the reach of today’s artificial intelligence technology.
Harnessing emotionally intelligent AI
As we at Phrasee have always said, investing in AI is only worthwhile when you have a problem to solve that AI can solve more efficiently than humans can.
Ergo, AI that is aware of the emotions of the humans it interacts with is useful only when it can perform a specified task more effectively based on the information that emotional awareness provides.
The question is: is emotionally intelligent AI useful?
Indeed it is.
In any situation where a human-AI interface is necessary, where AI needs to communicate with humans in some way, emotional intelligence can go a long way towards making those interactions smoother and more effective.
While we can’t claim that Phrasee’s AI is emotionally intelligent just yet, AI expert Dr. Neil Yager believes it is well on its way:
“Being emotionally intelligent aids communication, and communication is at the core of Phrasee. In its current form, Phrasee uses emotion indirectly: it tracks sentiment and the behavioural impact of different types of language. In the future, there will be a tighter feedback loop between emotions and communication.”
There are, however, other areas where that “feedback loop” is already breaking exciting new ground in the rapidly-evolving interface between human and AI.
This video showcasing Google Assistant making phone calls to businesses at Google I/O 2018 offers a glimpse into the potential this interface offers. Although it’s debatable what part emotional intelligence plays in such interactions, it is clear that a solid foundation is being laid for a future where AI-driven communication becomes the rule, rather than the exception.
From companion robots to AI Therapy, technological advancements are blurring the line the line between artificial intelligence and emotional intelligence in amazing and unexpected ways, and a future where artificial intelligence understands our emotions and emotional responses better than we do ourselves may be coming sooner than we think.