Emotional artificial intelligence. Understand and provide services.
Emotions are essential for humans. Practice has shown that artificial intelligence is as good at interpreting them as any qualified professional. This article will explain how emotional AI works, what data it uses, where it is applied and, essentially, what potential benefits it has for various businesses.
Emotions are crucial components of any human’s behaviour. Scientists define emotions as subjective states associated with assessment of internal and external stimuli which are expressed in the form of immediate emotional response. Understanding the emotional state of people is a key to interaction with them. That is why the best psychologists have always been good at interpreting people’s emotions.
Evaluation of emotional state can help businesses build up an efficient strategy of interaction with the customers, educational institutions — improve educational processes and make them more efficient as well, the entertainment industry — offer the personalized experience the client has been searching for. That is why the EDR (Emotion Detection and Recognition) market keeps growing incredibly fast. According to Mordor Intelligence, in 2019 the amount of Emotional AI investments was $17.2 billion and it is supposed to reach up to $45,5 billion by 2025.
How does emotional AI work?
Emotional states of a person are closely related to the body hormones level. Thus, Hugo Lövheim’s model compares basic affects, our emotional responses, which even babies demonstrate, such as joy, fear, grief, disgust, satisfaction etc., to the level of three neuromodulators — noradrenaline, dopamine and serotonin. This is a successful model for emotions interpreting!
Hormones affect different characteristics of the human’s state, from blood pressure and pulse to the moisture level of the skin and pupils size. Eventually, it is possible to get more precise information using sensors but many successfully implemented projects prove that we can rely on such secondary factors as:
- Intonation — the tone of the person reveals the emotional state;
- Facial expressions — minor movements of facial muscles usually reflect whole range of humans’ emotions;
- Eyes — pupils size, opened or lowered eyelids witness tension or relaxation, gaze direction reflects the characteristics of thought process;
- Body language — movements of hands, core, head positions, all that tells us about one’s inner feelings, excitement or relaxation;
- -Pulse and blood pressure — clearly reflect adrenaline and stress levels.
Thus, voice records and video analysis allows us to receive more secondary information about a person’s emotional state. Sufficiently trained algorithms are not only capable of identifying the emotional type of every person but also of tracking any change of the emotional state in real time.
In most cases you can get this data after politely asking permission to the user’s web camera access. As for pulse and pressure data, they can be easily measured by wearable electronics, which become more and more popular around the globe. Thus, even if you don’t know about it yet, you’ve got all you need to start working with emotional AI.
Emotional AI application examples
Examples of emotional AI implementation can be found in many fields and areas.
Gaming platforms
Working with games is one of the most popular and promising areas of emotional AI implementation. Gamers’ emotional state and the engagement during the game has a direct impact on their wish to continue the game, as well as on in-game content and advanced subscription purchasing.
Emotions analysis helps create customized offers and implement personalized players’ tracking. Under close integration of the emotional AI tool and the game there is a possibility to adapt the development of the game plot line to each user. For instance, psychological thriller Nevermind updates the picture depending on a player’s tension level analysis. The plot line becomes darker when the gamer’s stress level is high and vice versa.
Commerce and sales
Artificial intelligence allows to assess how a person perceives a certain commercial. For instance, several years ago such giants as Procter & Gamble and Welmart started to use the first digital price tags with commercials which adapt to users’ emotional state. A built-in camera sends the video to AI, which understands current emotion of the client and changes the commercial according to a special algorithm. As easy as it seems, this approach has helped increase sales by 10–40% as well as make users up to 3–5 times more engaged.
Education
China is one of the leaders in emotional AI implementation. Many Chinese educational institutions have been using emotions recognition systems in order to assess the involvement of each student. Not only may one get a bad mark for not learning the lesson, but also for being absent-minded during the class. Thus, the medea reports about the successful implementation of this system in one of the oldest schools in Hangzhou.
At Smartcoders we implement emotional AI in order to assess engagement and to understand the preferences of educational projects customers. Thus, students’ emotional state analysis is applied in Landscape Design School. In this field emotional AI helps assess students engagement and create really interesting learning experiences with personalized tracking.
Financial institutions
What could be more important than borrower’s emotional state analysis?
Knowing if the client is worried or confident, telling lies or being honest, can be much more efficient scoring data than Credit Report Bureau certificates and salary verification statements.
These technologies are not only reserved for some audacious fin-tech startups. For instance, one of the largest Russian banks Sber started testing similar technologies back in 2015. And in some big credit institutions of Societe Generale group special neural networks, analyzing customer’s emotional state based both on intonation and language, have already been implemented.
Security Services
Analytical systems integrated with video surveillance systems allow to spot potentially dangerous people in public places. Thanks to artificial intelligence security officers may receive alert notification in case of any suspicious behavior. Taking into account the fact that AI today is capable of spotting aggression, fear and other basic emotions with a high level of accuracy, security officers get the opportunity to react to any possible law violation faster.
By the way, a similar system has been used for several years by Chicago police in order to analyze data flow from more than 30 thousand cameras. The police officers receive an automated alert notification as soon as artificial intelligence detects a potential conflict.
Human Resources
Recruiting nowadays is a challenging task. Not only are there personnel shortages in the most advanced high-tech areas, but the Covid-19 pandemic requires to switch all interviews online only. Here emotional AI comes in handy helping recruiters get a lot of useful data even before watching the records.
Thus, American recruiter HireVue has been using a specially developed neural network to assess candidates’ state based on video interviews. AI analyzes keywords presence or absence, voice tone, body language and facial expressions. The score of the candidate depends on the compliance to the requirements set for the position. Recruiters of one British company receive AI comments after each interview. The report of the neural network points out how honest, curious, excited and confident the candidate was, how he reacted to the questions. This kind of analysis helps HR managers make more informed choices.
Emotional artificial intelligence is already available
Of course, there are still some limitations for emotional AI application. So far, the technologies don’t allow taking context into consideration, they are able to understand only basic emotions with an accuracy level of 70–80%. But even this turns out to be quite enough to solve current business tasks. By the way, the more you work with emotional AI the more precise analysis of your customers and users it provides.
In order to save time, together with Smartcoders you can get started on emotional AI today. Having a range of tested algorithms allows us to gradually connect you to the available data sources, using machine learning to solve specific business tasks. You can preorder a solution for your project or develop your project as our partner with minimal initial investments in AI, which is going to change your business.
