Can an algorithm help with existential angst?
Some of the latest progress in Artificial Intelligence is "Artificial Emotional Intelligence." This is "Technology [that] is prevalent in a broad way to learn, think, and act like human intelligence...Artificial emotional intelligence is a subset of artificial intelligence which refers to recollecting, recognizing and reacting to human emotions (Erol et al., 2020)."
There have been critiques leveled at applications for this technology like therapy chatbots, but this opinion piece had surprisingly positive things to say about WoeBot.
"After the growth of deep learning algorithms and neural network ap- proaches, emotion recognition captures high accuracy and better effi- ciency and inspires different perspectives in researchers. The major emotion recognition applications could be identifying attackers, potential criminals, or theft to prevent a crime. In the future, we are expecting large-scale testing of emotional reaction capture by facial, Speech, and text-based advanced machine learning algorithms and multi-layer schemes with high recognition accuracy rates for each sector of emotion recognition" (17).
However, this opinion piece does describe the—admittedly sometimes useful—gamification of mental health, and thus facilitating the use of programs to treat mental health issues. What is rather worrying is that to improve the algorithm it often needs real life data, and the mass collection of personal interactions between people and therapy bots, even with consent, has troubling implications when it comes to privacy violations and Skynet-level "crime fighting" oppression.
















