Just as in military operations, trust is essential to smooth communications and human-robot interaction.
Near the end of World War II, Adm. William “Bull” Halsey of the famed Third Fleet steered directly into a typhoon, losing 790 crewmen and 146 aircraft in the Pacific Ocean. Dr. Eric Daimler, formerly an innovation fellow with the Obama White House, shared this story with me as an illustration of the importance of trusting data. It turned out that Halsey’s mistake was listening to the weather report instead of his crew.
Six months later, Adm. Halsey again steered into a typhoon, but this time losing only six men and 75 planes. It was not until the postwar era, and especially after NASA launched the first weather satellite in 1960, that severe weather predictions started earning the trust of seafaring men and women.
Breaking down barriers between users and machines
Trust was the key takeaway from last week’s RobotLab discussion series on “Society 2.0: Understanding the Human-Robot Connection in Improving the World,” with Alexis Block of ETH Zurich and the Max Planck Institute for Intelligent Systems and Andrew Flett of Mobility Impact Partners. In a world littered with social robot failures, Block is a refreshing voice. The doctoral student’s research provides a useful guide in breaking down the barriers between machines and humans, relieving many of dystopian anxiety.
When Block first shared with the media her work of studying the effects of robo-hugs, many commentators were less than supportive, calling the innovation sad, dumb, or a waste of time. Reading through the numerous negative posts, it appears that the primary cry was the fear of a humanoid replacing the warmth of a living being. Block told a packed audience at SOSA’s Global Cyber Center that she is not discouraged and affably welcomes the criticism. Changing the mindset is actually the crux of her thesis.
Robots could actually help strengthen social bonds
To be clear, Block’s HuggieBot is a serious endeavor for people in need of human connections, such as children with autism and elderly dementia patients.
“We believe HuggieBot could one day be used to help separated friends or family — like college students or elderly people in nursing homes — feel closer to people who live far away,” she explained. “Also, it could be used to help people on the autism spectrum who have a difficult time experiencing physical contact from other people but for whom deep pressure touch is extremely beneficial.”
Hugs play a critical part to a person’s well-being, as each squeeze emits oxytocin, a hormone responsible for a person’s happiness. This “love hormone” is crucial for forming bonds between people, prompting the ETH scholar to question whether a mechanical act could elicit similar results.
“It is important to analyze people’s reactions to hugging robots because it’s possible that people could react very differently to a robot than how they react to another person,” Block noted. “In order to increase user acceptance of a robot, you need to understand how people feel about the robot and then design it, both physically and behaviorally, in the most socially acceptable manner for the majority of people. Otherwise, people just won’t use the robot because they don’t like it.”
This hypothesis has applications far beyond social robotics, as the findings could be utilized by integrators for deploying collaborative robots and autonomous vehicles.
The experiment started with comparing the preconceived notions of robot hugs with post-experience interviews. Block used a Willow Garage Personal Robot 2 outfitted in a variety of different exteriors and and squeeze timings.
“Our research found that robots should hug very similarly to how people do: they should be soft, be warm, squeeze us slightly, and release immediately when we nonverbally indicate we’re ready,” she said.
Block recorded levels of discomfort with long-duration hugs that wrapped cold metal arms around the human subjects. Today, she is building Version 2.0 with smaller Kinova Jaco arms that will be both slimmer and quieter.
“HuggieBot 2.0 will ideally be able to begin the hug motion when it notices a person is approaching it by using a depth camera,” Block said. “It will also have the ability to identify different approaching user arm positions and be able to use the appropriate reciprocating arm positions.”
A key aspect of the new study will be adding videos of loved ones to the screen of the robot to monitor reactions. “The new HuggieBot will have a screen that will be able to show either an animated face or a video message from a loved one,” said Block.
The improved HuggieBot will launch promoting “Free Hugs” next final exam season at ETH Zurich and the University of Zurich, providing overworked students with badly needed stress-relief.
Social robots to bounce back?
Block sees her machine as a bridge to eventually having social robots as a member of the family in the home. “I think it is natural to develop an attachment to these robots similar to what you might have with your own friends and family, so this could be another way to connect socially and physically to robots in the home,” the scientist said.
“I think given their high price point and limited abilities, these social robots just weren’t enough to convince users that it was a worthwhile investment,” responded Block. “As with any new product, when you make smaller batches, the cost per item is more expensive.”
“I don’t think these companies did anything wrong,” she continued. “If anything, I think it was a combination of the market not being ready to yet adopt the technology, especially because of the current climate of privacy concerns, combined with the limited capabilities at the high price point.”
However, Block said she remains optimistic for the future of consumer mechatronics, “because I see how excited people get in the mall when they see a Pepper just wheeling around. I think once privacy concerns are better addressed and we can provide a social robot at a reasonable price with a variety of capabilities — not just a tablet you can talk to but something that can physically and socially interact with its users in several different ways — then we’ll see a longer-lasting social robotics company.”
Block said she sees her contribution as more of helpmate than a novelty of an iPad on wheels. “Other humanoid robots have been created to help with more practical interactions with patients, like moving them from a bed to a wheel chair or an MRI,” she said. “HuggieBot sits in between these two: It is a humanoid robot that physically interacts with a patient but appeals to the emotional side of the patient.”
At the end of the event, I spoke with a colleague from a venture capital firm that focuses on female founders. Reflecting on this discussion, I found myself thinking about the first time Charles Babbage met Ada Lovelace in the early nineteenth century. Similarly to how Lovelace created the foundation for computer programming with her original thinking, Block could bring a unique perspective to a cold industry that longs for mass adoption. Her approach could not only forge trust with robots, but perhaps even encourage people to value one another’s companionship more.
“We view this technology as a supplement to human hugs, suitable for situations when requesting the comfort of a hug from another person is impossible, difficult, or uncomfortable.” Block contended. “We do not view this technology as an eventual replacement of human hugs or human connections.”
“As with any technology, the public needs to be educated not only on how to use the technology but also how to appropriately use the technology,” she asserted. “Just as too much television or social media is not good for human connections, relying too much on technology for a relationship is also not something we recommend.”