Can We Offend Androids and Artificial Intelligence Machines?

The uncanny valley of robots.

Source: Creative commons, Tiber Ergur

In the 1982 film Blade Runner, private eye Rick Deckard uses the Voight-Kampf apparatus to determine whether an individual is a synthetic human, or replicant. The device monitors eye reactions while Deckard asks questions to gauge the subject’s attitude and worldview. “What would you do if someone gave you a calfskin wallet?” he asks. As with Jules Verne’s Nautilus submarine or George Orwell’s newspeak, modern reality is increasingly making fictions like these seem more like a documentary.

The Uncanny Valley Effect

Neuroscientists at the Cambridge University recently scanned participants to quantify their neural reactions to images of humans versus robots, accompanied by ratings that measured subjects’ likability and humanness. Their brain activity altered as the line between robot and human became more blurred, specifically activity in the amygdala which is associated with anxiety, fear, and negative valence. The result is called the “uncanny valley effect,” meaning the way we get creeped out by a human-looking robot whose artificiality is nonetheless noticeable.

Spitting out humanoids like a knitting machine.

Source: Neogaf, Common commons

Earlier I wrote about how the HBO show Westworld depicted love as a spectrum and suggested that even if one’s objects of desire were built from silicon, the romantic feelings it elicited would be genuine. In other words, could a TV show teach us to love robots?

The android hosts of Westworld are designed to seem as human and likable as possible so that visitors feel unencumbered to act out their unacknowledged desires, whether that is adventuring in the mountains or falling in love by a river. Yet while Westworld and Blade Runner are works of fiction, the Cambridge research suggested that we are at least starting to consider what it will take to create artificial beings with whom we can connect.

Sex Toy or Calculator Companion?

Will androids exist as commodities meant to placate flesh-and-bone humans or will they be a new means to replicate life in a way as easy as picking out a do-it-yourself item at IKEA? Regardless of one’s attitude, the uncanny valley effect highlights a primal skepticism toward artifacts that pretend to be human.

Being friendly with human-seeming robots in the future will require rewiring our brain. This will either be done by engineers mitigating the uncanny valley effect through more sophisticated software and algorithms, or by the populace in general learning to not let the uncanny valley bother us so much.

Comment to Dr. Cytowic at neuroman@gwu.edu or to ask for his low-frequency newsletter and a copy of “Digital Distractions: Your Brain on Screens."