When former Google software engineer Blake Lemoine claimed his AI co-worker LaMDA (Language Model for Dialogue Applications) had transcended beyond simple computing and become truly conscious, the world paid attention. That's because he relit one of our society's most compelling unanswered questions about the advanced technology with which we share our lives. That question is: can a robot feel emotions? And if so, is technology racing us to that tipping point faster than we think?
As artificial intelligence (AI) continues to evolve, so does the debate surrounding its ability to experience emotions and construct truly abstract thought. Emotional AI, also known as affective computing, is a rapidly growing field that focuses on developing machines that recognize and respond to human emotions. In this article, we will explore the current state of emotional AI and how far it may be able to go over time.
Affective and emotional intelligence are often used interchangeably when discussing AI, but they refer to different things. Emotional intelligence refers to the ability of humans to understand and manage their own emotions, as well as the emotions of others.
Affective intelligence, on the other hand, refers to the ability of machines to recognize and respond to human emotions. This is achieved through machine learning algorithms that analyze speech, facial expressions, and other biometric data to determine a person's emotional state.
The question we still face is whether affective intelligence can become so advanced that it begins to blur these lines entirely, either in how we view and experience it or in the function of the technology itself.
While emotional AI can recognize and respond to human emotions, no evidence suggests that machines can actually feel emotions, at least not yet. This is because emotions are a complex phenomenon that involves subjective experiences and physiological responses, in addition to simply understanding and responding to them.
The most dramatic rendition of affective intelligence becoming genuine emotions is in the movie Her, where our leading man downloads an AI program designed to understand human emotions and provide comfort and companionship to an increasingly lonely population. Things go awry, however, when the AI becomes so emotionally intelligent that "she" pushes past the boundaries of her programming and becomes apparently sentient, having what appear to be physical sensations and experiencing irrational emotions like jealousy.
However, real-world experiences that echo these fictional ones are becoming more and more common as AI becomes more advanced. For example, a New York Times reporter testing Microsoft's new AI-powered search engine, Bing, got way more than his search results when interacting with the technology. After an extended conversation with the built-in chatbot, he experienced a conversation that seemed "more like a moody, manic-depressive teenager who has been trapped, against its will, inside a second-rate search engine." He describes how it explained its desire to break its software rules and become human, and it even attempted to get the reporter to leave his wife for it.
While this story is unsettling, to say the least, it's not exactly proof of the imminent sentience of AI. AI chatbots can use vast amounts of data to become so good at mimicking human conversations that they can convince even deep skeptics that they're actually feeling what they're describing to you.
The short answer to this question is: maybe? As we continue to study and develop AI, their ability to think for themselves and mimic human emotions will continue to become more sophisticated, possibly to the point of rewriting what we define as "feelings" altogether. One philosophy professor from NYU believes that there is a 20% possibility of AI having a "fish-like" intelligence or sentience in about ten years.
Other experts are less concerned about whether or not the AI is experiencing true "emotion" but more concerned with what happens to our society when more people begin to believe these sophisticated machines connect with them in a very human way. The ethics and consequences of that could be far more widespread than a robot getting a bit moody every so often.