AI systems and devices will soon recognize, interpret, process, and simulate human emotions. A combination of facial analysis, voice pattern analysis, and deep learning can already decode human emotions for market research and political polling purposes.
Source: Kleber, S. (2018). 3 Ways AI Is Getting More Emotional.
There are currently 3 categories for what is known as emotional AI:
- Systems that use emotional analysis to adjust their response e.g. routing consumers to appropriate services (including human operators) based on their emotional status.
- Systems that provide a targeted emotional analysis for learning purposes e.g. bracelets that monitor a user’s emotional status and alerts them when they’re in a heightened emotional state and so influence decision-making; may also have utility for people on the autism spectrum, providing insight into the emotional states of others.
- Systems that mimic and ultimately replace
human-to- human interactions e.g. conversational user-interfaces that take on various social or professional roles in their interactions with people.
Of course, there are concerns with these kinds of systems that are “playing on the edge of ethical acceptance”. The most important step moving forward won’t be to get the emotional intelligence of the AI right, but rather to make sure that software designers have the emotional intelligence to understand the impact of their systems.