When artist Agnieszka Pilat ventures out with Boston Dynamics robot ‘Spot’, she always wears yellow, making the human-robot team present like a pair of Minions.
Colour-coordinating is a deliberate ploy by Pilat, designed to make the robot more approachable, to inspire trust and curiosity, and help overcome people’s fears and misconceptions.
“It puts people at ease instantly when they see there’s a human connected to the robot, not that the robot is walking by itself,” she tells Cosmos.
When she’s outside walking with Spot, children in particular can start off a little bit afraid. “But that goes away fast, because their curiosity is so huge, they just can’t stop themselves.”
Pilat’s aim as an artist is to demystify the technology, so she frames Spot for the public as child-like, “kind of awkward” and “learning a lot”.
Unlike artificial intelligence or generative art which lives online, robots have to navigate the physical world, Pilat says. And it’s their mistakes and imperfections in doing so which brings them closer to us as humans, she says.
Pilat and her robot collaborators are on their way to Melbourne this year as part of the National Gallery of Victoria’s Triennial, coinciding with the Australian Government’s development of the first National Robotics Strategy.
‘Trust and inclusion’ is one of four central themes highlighted in the robotics strategy discussion paper released earlier this month. Increasing public understanding and ensuring trustworthy, ethical and responsible use of robotics is crucial to the success of the industry, the paper says.
Professor Mari Velonaki is an expert in social robotics and the founder and director of the Creative Robotics Lab and the National Facility for Human Robot Interaction Research at University of NSW.
She says, deploying robots as artists or guides in cultural settings, like galleries and museums, offers a more playful setting for people to interact with the technology, allowing researchers to explore and observe people’s responses and interactions across a diverse range of age groups, identities and cultural backgrounds.
Her research investigates how people interact with, and respond to social robots (robots designed to engage with people using social behaviours) – in different contexts like homes, aged care, hospitals, education or other workplaces.
Velonaki has been working in the field since the early 2000s and has seen the focus of the robotics sector shift from engineering and technical design, to a much greater emphasis on understanding the human element.
She welcomes the openness of the conversation around the National Robotics Strategy, and hopes it will ultimately lead to greater collaboration between industry and academia, and more multi-disciplinary design practices.
She says, more dialogue between ethicists, social scientists and engineers from the outset will also enable better, and more interesting designs.
“If you look at big companies, there’s very specific stereotypical robotic forms. We don’t explore material as much as we should. […] We don’t have the robots that we deserve for our society.”
Building trust, acceptance and understanding of robots requires people to experience them in the real world, Velonaki says. Doing so helps people understand what robots are – and aren’t – capable of, offsetting ideas based on science fiction like fears about killer robots, or evil machines.
She says for some topics, like data and privacy, there needs to be a much broader public conversation. She describes this a double-edged sword for robotics – the more information a robot takes in through its sensors, the more efficient and useful it will be; yet on the other hand, the more privacy concerns people may have.
Evaluating new technologies and products before they are commercially produced is also crucially important – for both industry and researchers – to understand whether certain designs evoke attraction, or repulsion in potential users, she says.
Classically trained as an artist, Pilat works with technology both as the subject of her portraiture, and increasingly as a collaborator. She has worked as an artist-in-residence at technology companies including Boston Dynamics, Agility Robotics, and SpaceX.
For the NGV exhibition, Pilat is training three robot dogs to paint ‘by themselves’. She hopes it will spur a stronger reaction of curiosity or even discomfort, by creating the impression that the “there might be some conspiracy and they’re communicating”.
Alongside the exhibition, RMIT researchers will be recording how members of the public respond to the robots in the gallery space.
“Seeing a robot creating art, in Melbourne’s premier gallery, challenges our ideas about what a robotic future could look like,” says RMIT Chief investigator Brad Crammond.
As the robotic dogs manoeuvre their sticks daubed with oil paints onto the canvas, the robots might look like they have their own agency, Pilat says, but they really don’t.
“The instructions come from me,” she says.
That doesn’t mean the task is simple. She explains a principle called “Moravec’s paradox” which means “in robotics difficult is easy, and easy is difficult”.
“For us human beings, mathematical equations, and huge sets of data are very, very difficult. But eye-hand coordination, movement, environment, is super easy.
“In robotics it is the opposite. So a robot, which is driven by algorithms can answer a maths problem instantly. But it takes a lot of computation power to actually do sensory perception in the environment.”
Read more on the National Robotics Strategy