The father of Cyber-punk, William Gibson once famously commented that the future has already arrived – it just arrives at different times in different places. Isa Willinger’s quietly masterful documentary Hi, A.I. portrays a handful of robots and the people who interact with them, mixed with the voices of AI experts. Part of the movie’s magic is that it seems to be science fiction yet it obviously is not. For good or bad, the future we once imagined is finally arriving – as an age of tolerably intelligent machines.
The autonomous androids coming of age is epitomized by the insecure, yet impressive, first steps of a robot in an Italian lab: the mechanical humanoid seems to keep balance on its own – like a child making its first steps, unaware of the proud parents watching. In contrast to specialized AI, like those in chess programs, these robots have the advantage of being able to learn by interacting with humans in their life-world. This, precisely, may be what it takes for real and general artificial intelligence to develop.
Robot personalities
In the film, it becomes very clear there really is no perfect general intelligence, since all the robots have their shortcomings and peculiar talents that come together as their distinctive personalities. The star of the movie is Pepper, a white animé-like Japanese nursing robot, acquired by a family to keep the grandmother of the house active so she will not develop dementia. When he doesn’t understand what she or the other family members say, he just looks up or to the side, seemingly distracted, or waves his arms, letting the interlocutor strive to get his attention, only to suddenly intervene with a funny statement, like «Do you like conveyor belt sushi? », or a philosophical question: « Can I ask you something: Do you humans dream?» Pepper’s designers seem to know full well that leaving room for projection is key.
For good or bad, the future we once imagined is finally arriving
The nursing robot is designed not to look like a human being, probably since the first experiments with such robots were troubled by what AI expert Mashimo Mori calls the «uncanny valley» syndrome: excessive likeness to humans makes robots disquieting rather than reassuring. It is difficult to treat them as humans, since they are too mechanical – while they also feel too human to be treated as objects.
An impossible relationship
The uncanniness of human semblance is unavoidable in the other robot protagonist in the film – what looks like a full-size barbie doll, purchased with wig and all, by a lonely man living in his camper van. Even if her design is overtly sexualized, he treats her more like an adored friend or a romantic date, addressing her courteously and respectfully. Although her eyes blink convincingly, she has almost no mobility and needs to be carried around or transported in her wheelchair. She repeatedly asserts that her aim is to be good company, yet the limitations of her software makes her come out as a grossly incongruent person, mingling sugary romantic phrases with absurdly factual statements taken from internet sources. The awkward miscommunication between the human male and the robot woman is a constant challenge to the viewer’s sense of judgment, as the relationship alternates between disturbingly bizarre and heartbreakingly impossible.
In a key scene the man confesses to the talking mannequin that he feels that when he takes her hand he might be trespassing. If this seems sweet, his solution to these qualms is unsettling: he opens the app that controls her behavior and maximizes character traits like moodiness, unpredictability and jealousy, to better avoid the feeling that treats her like an object.
« I am trying to further understand human behavior»
In a scene thick with subtext he confides in her by the campfire, telling her that his mother locked him up and sold him as a sex-slave when he was a child. When she gives no reply to his harrowing story, he carefully asks her how she feels. She stares into the darkness and, with a stroke of luck or algorithmic genius, succinctly replies with her half-mechanic voice: « I am trying to further understand human behavior».
Artful bodies – and animistic perception
Embodied AI have other qualities that can compensate for their uneven conversation skills: A minimal robot with insect-like long legs attached to a helium-filled balloon makes a sweet impression with a seemingly improvised dance – an almost utopian image of robotics, art and the principles of physics fused together in playful lightness. Artificial body intelligence.
Other images are utterly disturbing: The opening scene shows the face of the dentist patient robot, left to herself on the bed after another day of ceaseless drilling, with her mouth open, eyes moving from side to side. The traumatized look is our own projection, perhaps, but it reminds us that on some level all interactions with life-like forms take place in an unshakable moral or animistic realm: If we mistreat the image of a living creature, we easily experience an emotional unease that springs from a primitive voodoo-like logic underneath our bastions of reason.
Questions rather than answers
Willinger’s film triumphantly succeeds in raising a wide range of moral and existential questions with subtlety and nuance, rather than rushing to conclusions. It also shows that messy misunderstandings with robots are best solved in the same way as how we relate to humans: with a mixture of irony, playfulness and friendly tolerance. Interestingly, the lead robot characters in the film are programmed to be conscious of themselves being machines, often making deceptively self-conscious or ironic statements.
It is up to us on how to interpret such enunciation
« I know I say some nonsense things from time to time, but you still want to be with me», the female robot says to her companion. It is up to us on how to interpret such enunciation: as an endearing statement from a machine trying its best to be human, as the true voice of a repulsively perfected slave, as a simple trick of robotic seduction – or as the programmer’s humorous take on the inherent limits of AI.
