Is AI Capable of Empathy?

Your Uber arrives at the exact time estimated by the app.
The smartspeaker remembers the last type of coffee you ordered for your Keurig press, and when to order it next.

Artificial intelligence creates these possibilities, where “machine learning” trains computers over time with data/examples to complete tasks. 

It seems like computers are getting to know you. Catered towards your needs, you feeling a sense of amusement and comfort. At the very least, they’re making you feel.

How far do we have to go until they could feel, and understand you for more than someone to recommend Netflix shows to? AI and empathy is an intersection where we can discuss the extent of AI in the near future. 

Facial Recognition, a Path to AI Empathy?

Empathy is the capability to understand and share the feelings of others, and it’s been a uniquely human thing. There seems to be an area in machine learning that is similar to that department of “human-only” skills: Facial recognition. It’s possible through computer cameras to capture facial features, and machine learning algorithms to interpret the emotions behind the expressions. 

A data set with over 7.5 million faces over 87 countries is making facial recognition eerily accurate. Its current use cases are to detect our emotions for market research, identify impaired drivers on the road, or for medical professionals to check in with the wellness of their patients.

However, it’s… not a stand-in for empathy. Yes, it can read how you feel, but not the circumstances that drive why you feel. Empathy only happens when you try and step into the shoes of another person. You need to feel emotions from a context away from your own. 

With the current state of AI, they aren’t capable of what we traditionally know as empathy. But can we emulate it? Will we need to?

An AI-Powered Chatbot Possibility: Mental Health

We can’t make chatbots understand our emotions, but we can make them simulate empathy. We can mimic a human conversation and add to a script to make it more human-friendly (as illustrated in this article). 

Chatbots may be perceived as empathetic, without actually being empathetic. Mental health is an area where AI could see incredible potential. There is a shortage of mental health professionals and resulting cost is a barrier to patients. Could AI help identify depressive symptoms through behavioural pattern recognition? As a 24/7, cost effective tool to provide immediate support, AI-powered chatbots have a potential niche.

For now AI can only answer generic and basic mental health advice. AI will not have emotional intelligence as we know it. If the AI isn’t trained with a diverse dataset, potential bias may occur. Other limitations are privacy concerns, and the lack of safety regulation for AI to be a 100% stand-in for professionals. However, if introducing an AI chatbot can alleviate accessibility issues as an initial touchpoint to prevent harm, or schedule and encourage individuals to seek help – those are worthy things we can look for.

AI Today 

When we talk about AI, there’s discussion of whether or not they’ll actually be capable of taking our jobs.  We know that humans have the ability to feel, be empathetic, and form new answers – and that will separate us from the robots in the time being. AI may not be taking our jobs right now, but they’re currently used as an efficient solution to aid business processes and workflows. 

To add on to the earlier list, AI enables us to do many things already. It’s what opens our phones when we use face ID, helps you send an email or message, personalizes your social media feeds. AI-powered chatbots now exist, using machine learning to better execute a task that you’ve assigned them to do. They’re increasingly present in businesses to automate tasks, and an effective point-of-contact with customers. 

Chatnels is an AI-powered chatbot platform that can bring digital assistants into your workflow. Start with Chatnels today.