Robots in Love

Nowadays, colloquially labeling smart robotics as ‘artificial intelligence’ became regular practice because it’s convenient, and it’s almost like a buzzword that screams progress.

Scientists are a bit more sceptical in throwing around that word, with some claiming that true artificial intelligence doesn’t yet exist, calling it augmented intelligence at best. Sophia might throw out a couple of fun one-liners, but there’s nothing authentic about her thought patterns. Let’s say if Sophia was your girlfriend, she wouldn’t cheat or get jealous, and she’d only love you to a certain degree since her emotional capacity is limited to pre-arranged code. We still haven’t developed advanced synthetic intelligence yet, one that is able to create instead of imitate, and that’s why the concept of love between robots and humans is still arguable. There was an episode in Black Mirror where a woman recreated her deceased boyfriend, only to find that android Ash lacked small details that made him unique. For example, the android follows Martha’s orders, whereas Ash would have questioned them. The robot was able to acquire new features, but only through imitation, and that’s why Martha eventually distanced herself from the android. A stark contrast to Sapper Morton, a replicant robot from Blade Runner 2047 who cared for all life, gentle and even righteous, yet still hunted down by humans as a ‘rogue skin job’. 

Actually the foundational rules regarding robotic technology have already been set in the 1940s when writer Isaac Asimov came up with his Three Laws of Robotics that were meant to keep robots from harming humans. When the idea of AI was conceived in the 1950s, people largely overestimated its abilities, with computer scientist and MIT AI lab co-founder Marvin Minsky boldly claiming that “within a generation…the problem of creating artificial intelligence will substantially be solved”. There was so much hype around artificial intelligence that it barely had any opposition. But there was Hubert Dreyfus, an outspoken MIT philosopher who argued that computers don’t have any cultural background, childhood upbringing or consistent memories that play a crucial role in brain development, therefore their intellectual capacity is formulaic at best, lacking common sense knowledge which cannot be incorporated in a computer program. 

For better or worse, Hubert Dreyfus’ claims still live up to this day. We have yet to see AI that can exist within its own realm of thought. Pioneering engineer Judea Pearl, known for developing probabilistic AI in the 80s, is convinced that truly intelligent machines should understand the relation between cause and effect, meaning that a sapient robot should independently figure out that smoking causes lung cancer, not just correlate between smoking and cancer. They also need to attempt answering questions  such as “Why did I do this?” or “How could I make this better?” According to Judea Pearl, the goal is to replace reasoning by association with causal-reasoning, teaching robots to deduce through observation. A combination of data and causal reasoning could spark a mini-revolution in AI, with systems that can plan without playing imitation games, even leading to free will in robots. Nonetheless, Mr. Pearl is impressed with achievements that are already made, but he encourages scientists to think forward because deep learning is stuck in a curve fitting rut. Cognitive scientist Daniel Dennett sees only two scenarios in building emotional robots. It’s either programming AI to act like it’s in love (faking it) or making it think in a non-linear way, just as a human brain would. The ultimate goal would be a thought process that relies on a mesh of interconnected thoughts, sort of like a neural democracy. Only then can the possibility of robot-human romance ever be considered. 

Or maybe love is a mistake that has nothing to do with pure reason? Vytautė Žilinskaitė’s The Robot and the Moth (1978) describes Dondon, a robot that falls in love with an insect, resulting in programming errors caused due unknown emotions. 17th century philosopher René Descardes would have cheered as he was a staunch believer in mind-body dualism, convinced that intelligence is perfect only when it’s free of emotions. That kind of notion existed for a long time among Western intellectuals, right up until the early 90s when neuroscientist Antonio Damasio challenged the notion that reason and emotion are radically opposed, claiming that good reasoning is reliant on the capacity to experience and regulate feelings. Damasio’s neurological data suggested that without feelings (and their bodily accoutrement), cognitive capacities were surprisingly blunt and inept. To put it in simple terms, one could not be cognitively smart without also being emotionally capable since there’s a mutual relationship with every organ, placing the body directly within the chain of operations that generate reasoning, decision making, social behavior and creativity. Damasio argued that “emotion, feeling and biological regulation all play a role in human reason.”

It’s still early to say whether such androids are the right way of solving society’s psychological setbacks. But what we do know is that the seeds for human-robotic relations have already been planted. However, even if we did develop conscious artificial intelligence, the question still  remains whether we’d be able to sympathize with non-biological lifeforms. Isn’t that a fundamental obstacle in understanding one another? In Fallout 4, a post-apocalyptic RPG game, there was a dramatic standoff between two soldiers, one of them appearing to be a robot (he was unaware of it), and the other one, Elder Maxson, saying, “Danse isn’t a man, it’s a machine… an automaton created by the Institute. It wasn’t born from the womb of a loving mother, it was grown within the cold confines of a laboratory. Flesh is flesh. Machine is machine. The two were never meant to intertwine. By attempting to play God, the Institute has taken the sanctity of human life and corrupted it beyond measure.”

Elder Maxson refers to similar problems that Hubert Dreyfus mentioned. For people it could be difficult to love someone that’s not biologically alive, meaning that they haven’t experienced physical or mental development leading to an adult life. No traumas, no regrets. No risk of getting sick. No death? The question is how do you grow old with someone that’s self-aware yet doesn’t experience aging? Our collective consciousness might not be equipped in dealing with androids that don’t experience vulnerability. However, Tõnu Viik, a professor of philosophy at Tallinn University, offers a  way out of this thought loop by realizing that vulnerability comes in many different forms. Maybe a self-conscious android has fear of having its memory being wiped, or even an acute phobia of non-existence. 

In a 2020 paper, Viik claimed that emphasizing with uniquely robotic emotions can expand our understanding of vulnerability as a whole, going beyond biocentric/ anthropomorphic notions of what fear is, and enriching ourselves as human beings. He claims that “if these assumptions could be put aside, loving synthetic androids would make us better persons and expand our notion of love rather than undermine it.”

A 2015 study offers evidence that humans are at least able to empathize with robot pain, showcasing an experiment in which scientists used electroencephalography to observe human/robot photos in painful situations. Results suggest that we empathize with android pain in similar fashion to humans, though the beginning of the top-down process did show weaker empathy towards robots. Scientists attribute the initial lack of empathy to human inability in adopting robotic perspective. 

This could explain why a 2020 study showed that a robot expressing vulnerability makes humans more comfortable than ones without personality, showing positive impact in teams that have a machine displaying human-like qualities. Research suggests that people gravitate more towards robots that are anthropomorphic, meaning those that look familiar…human, so it’s reasonable to assume that humans could romance anthropomorphic androids. The question remains whether robots can love us back.

Giedrius Alkauskas, Senior Researcher at the Faculty of Mathematics and Informatics at the Vilnius University imagines a world with robots acting as our psychiatrists.