Why AI innovators want to know how you’re feeling

 

What if a robot could tell how you’re feeling before you can? If an AI algorithm was so in tune with human nature it could instantly detect whether you’re contented, anxious, excited?

That’s not science fiction. It’s the power of emotional AI - a type of artificial intelligence capable of recognizing and responding to human emotions. It could have countless applications in business, from sentiment analysis (analysing how large numbers of people are feeling about a public message) to personalised travel recommendations, to more effective customer service.

To see the value of emotional AI we first need to understand emotional intelligence: an individual skill linked to self-awareness, social awareness and personal management.

  • Self-awareness focuses on knowing your own emotions and those of other people.
  • Social awareness is the ability to understand others’ emotions and their relationship with others.
  • Personal management, on the other hand, refers to being able to regulate one’s own emotions and behaviors in order to achieve goals.

    People with high levels of emotional intelligence are usually good at solving problems, cooperating well with others and communicating clearly. Around 80% of vital leadership skills are based on these abilities.

    For business, emotional intelligence means understanding your customers. Knowing how you’re making them feel. Emotional AI provides that intelligence. It’s an advantageous tool because it can create more realistic customer experience models, including to anticipate customer needs, achieve product- market fit and develop marketing and communications strategy.

And the applications of emotional AI extend far beyond business. The use of such technology has already been investigated in relation to elderly care, particularly with robots. In one study, a group of participants were asked to help an elderly woman complete her task. Those that interacted with an emotionally intelligent robot perceived it as more helpful than those that interacted with a robot without emotional capabilities.

Here’s just three of the ways emotional AI could make a positive impact in the world:

• Education: Children show emotion more readily than adults, and that presents an opportunity to personalise their experiences in the classroom. Prototype learning software, powered by emotional AI, can adjust a task depending how it’s making the child feel; reducing the difficulty, for example, if the child is exhibiting frustration.

Children with autism, meanwhile, tend to struggle with recognizing the emotions of others and themselves - another potential future application of the technology.

• Fraud detection: Algorithms may be able to help humans overcome a major flaw faced by insurance businesses: knowing whether somebody is being honest. Emotional AI could be used to detect signs of untruthfulness when customers are making insurance claims, so the issues can be checked in more detail if necessary.

• Autonomousvehiclesanddriversafety:Carmanufacturershavealready developed the technology to identify when drivers are stressed - Ford, for one, debuted a prototype in 2018 - and soon, the tech will be fitted to commercially available vehicles. Emotional AI can detect when a driver is at higher risk of an accident - if they are stressed or tired, for example - and warn them accordingly.

But there’s a problem. Emotional AI isn't always accurate. It’s difficult for computers to consistently identify emotions accurately in photos, videos or spoken words. If somebody laughs as they tell you they’re unhappy, could an algorithm correctly interpret the nuance that humans naturally understand?

If somebody looks annoyed, but they tell you it’s fine, does the algorithm know that things probably aren’t fine? It’s unlikely.

Getting it wrong could have serious negative implications for individuals. If, say, the algorithm incorrectly interprets somebody’s emotions during a job interview one of the most commonly cited applications of the technology - it could derail the individual’s entire career.

As with many forms of AI, inherent bias is another in-built problem. AI developed largely by men, for example, may deliver inconsistent results for women. Businesses will need to determine how they want to address these biases if they're present in their AI system, either by incorporating human oversight into the process or by programming new biases into the system.

Both opportunities and challenges, then, lie ahead. It’s clear that emotional AI can solve real world problems and improve peoples’ lives, at scale and at relatively low cost. If innovators can address the ethical issues and make sure the technology becomes a force for good, the potential is almost limitless.

Ready to harness the power of artificial intelligence? SBM can help. Get in touch today to learn more.