Sci-fi tales are inherently based on the possibility that one day, a real person and a machine will develop a genuine, authentic, human-like liaison.
There are tons of movies and novels depicting the complex interactions between people and sentient-like machines, telling us about their mutual interests, their impossible love stories, their friendships revolving around the same goals, or dystopic stories of rebellion and vengeance. This storytelling is based on the premise that humans and machines are able to share feelings.
An Anthropocentric Mistake
We are driven by a cultural need to assign names, purposes, and even souls to inanimate objects to better relate to them. This tendency begins in early childhood, as we play with plush toys and pretend they are alive. Even though we know that robots or AI-generated creatures consist of circuits and bytes, our natural inclination to anthropomorphise them persists. This involves giving them human-like features in appearance and behaviour, which appears to be essential for enhancing our interaction with them.
The concepts of androids and mechanoids have existed since the early 1700s, and they have quickly evolved since then, reaching new peaks of realism over the past few years. Bionic people made of cyberskin and AI-fuelled brains could actually be among us; at this point, we are able to build them.
Visualizza questo post su Instagram
AI generated androids shows how biased is our concept of robot, as they have metallic body features but a high sense of style that humanise them
Nonetheless, our homes are not filled with lifelike cyborgs but with robots that are not designed to resemble humans. Still, they do have voices similar to ours (think of HomePod) or move around like pets (like Roomba). The intention behind designing them without anthropomorphic features is to ensure that we don’t grow attached to them or develop actual feelings for them. It’s important to remember they’re here just to serve us.
In less than a decade, personal assistant bots will outnumber us. Their natural language features are already so lifelike that we could be easily tricked into thinking they’re human, raising a huge ethical problem.
Their appearance, though, will prevent us from believing they’re actual people. That’s why product design for human-robot interaction is key to improving our experience with them.
Looking for Faces Everywhere
Pareidolia is a very human tendency to look for faces in inanimate things. It’s a biased behaviour that shows in many ways that we are wired for human connections. Put some googly eyes on a sock, and it will instantly become a rabbit.
Seeing faces everywhere is a natural habit. People have always had this funny habit of seeing human-like traits in things that aren’t human at all. We do it in stories, myths, and art all the time, giving animals, objects, even the weather, personalities and feelings just like us. And you know what? Faces play a big part in all of this.
From the moment we are born, we are naturally inclined to focus on faces. Faces are crucial in helping us comprehend other people’s feelings and how we should react. It’s no surprise that objects with features resembling faces attract our attention. Product design largely benefits from the intentional use of Pareidolia, adding subtle facial cues to everyday objects.
Visualizza questo post su Instagram
Building can stare at you! It's the effect of Pareidolia in action
Research shows that most people see human faces in car fronts and think that most car design styles show angry expression.
In 1997, the movie Flubber featured Robin Williams as genius Professor Philip Brainard. In the story, his flying robot assistant, Weebo, develops a crush on him. Weebo is shaped like a metal flying saucer with a nice little monitor and a cute, flirty female personality. It understands the importance of appearances and tries to seduce him by creating a holographic version of itself resembling a beautiful woman with long hair in a silk nightgown. This portrayal may seem a little stereotypical, but it accurately reflects the criteria for attractive characters for the late 90s.
This heart-warming vintage movie shows how difficult it is to overcome such an unbridgeable gap between being an actual human and being a machine trying to be one. Sci-fi has come up with interesting ideas over time. Giving human appearance to robots has always been key.
The Uncanny Valley Trap
From RoboCop to WALL•E, facial expressions are a key feature that makes us empathise with machines. Nonetheless, there is a huge risk that we will sink into the Uncanny Valley.
The Uncanny Valley phenomenon, unveiled and examined by robotics professor Masahiro Mori in the 1970s, posits that as a robot’s appearance approaches human likeness, people tend to respond more positively and empathetically. However, there is a point when the resemblance becomes so close to a human that our reaction quickly turns to repulsion. This shift stems from our tendency to have a negative emotional response to an inanimate object that, in its endeavour to mimic humanity, falls short. In essence, as the author suggests, we lose our sense of connection and affinity towards the robot, making it appear uncanny. Yet, if a robot’s appearance becomes indistinguishable from a human, our emotional response becomes positive again, approaching the levels of empathy experienced in human-to-human interactions.
Visualizza questo post su Instagram
Figure 01 made by Opan AI Robotics company is powered by GPT and is the most advanced humanoid robot ever made, until now
According to this theory, we are more likely to empathise with android twins Walter and David from Alien: Covenant because they are virtually identical to human twins. Additionally, their ability to speak with different accents makes them even more realistic. Other examples of androids we can catch feelings for include Maria from Metropolis or the Tin Man from The Wizard of Oz, with their shiny metal bodies.
Again, actor Robin Williams gives us an interesting exception for this staple theory when he played a robot in the movie Bicentennial Man in a 200-year storyline transition from a robotic appearance to a more human-like one. We like him anyway because of his personality: the Uncanny Valley theory doesn’t kick in because we can see through his appearance and deep into his soul.
Maybe a better future for human and robot interaction is not entirely linked to our ability to craft a perfect robotic human replica, a virtual clone, or a lifelike personality. Or making neural nets so perfect that we can chat, laugh and cry with them as we would with a real friend or a lover.
Perhaps improving the design, technology or features of robots and disguising them as humans could be one way to enhance our relationships with the many bots that are already part of our lives. However, we should explore more possibilities, at least until better solutions are found.
Gaia Giordani
Generative AI explorer and New Media Communication expert