The COVID-19 pandemic has affected every aspect of life for people around the world, but reactions have varied in terms of individual behavior, governmental policies, and attitudes toward technology. As the debate continues over whether or not automation will replace employment, a new survey found more reason for hope in a future where societies can benefit from robotics and artificial intelligence.
Interactions LLC surveyed 1,000 people across the U.S. and found that consumer sentiment was generally positive toward AI and robotics. The Franklin, Mass.-company worked with SurveyMonkey on the “Comfort with AI in a Post-COVID World” poll, which was conducted on April 28 and 29. The survey asked questions about how people felt about automation in customer service, grocery stores, healthcare, privacy, and self-driving cars in comparison with before the novel coronavirus crisis.
“As the coronavirus continues to have a radical impact on business and society, it has ushered in a fresh urgency for technological progress,” stated Jim Freeze, chief marketing officer of Interactions. “It’s well understood that AI can play a powerful role in operations. At the same time, it’s crucial that consumer comfort is kept front and center as organizations navigate the ‘new normal.'”
The need to reduce direct human contact for social distancing while still delivering services such as food have led to a better understanding of the roles that robots can play in healthcare and the supply chain. However, robotics designers and vendors should note that significant skepticism persists.
Survey results find comfort growing in retail
Of the survey respondents, 45% said they would prefer to have their inquiries handled by AI if it would be faster, and 21% indicated that they are now more comfortable having a full conversation with an AI agent. Previous studies found that as much as 87% of consumers preferred to deal with human representatives.
Thirty-three percent of consumers polled said they are now more comfortable with robots in grocery stores, but 24% said they felt less comfortable. Robots are already cleaning, taking inventory, and performing other customer-facing functions at retailers and restaurants.
“We didn’t ask survey participants to compare different types of robots, but we can look at how the comfort level differs across the AI applications addressed in the survey, like delivery robots and robots in grocery stores,” Freeze told The Robot Report. “One finding that stood out to me was the fact that 1 in 4 people feel so uncomfortable with robot deliveries that it would actually deter them from making the purchase in the first place.”
Privacy and using AI to track infection
As institutions rush to develop new ways of tracking people infected with COVID-19, 31% of those surveyed said they would be more willing to share personal data, while 22% of Americans were less willing. In addition, 47% of respondents said their willingness to share such information has not changed, with 25% already willing to do so, and 22% not willing. Smartphones and Web apps already collect a lot of personal data; so how are robots or AI different?
“The personal data that’s collected from smartphones and Web apps could be used to train machine-learning algorithms that identify patterns and high-level trends,” Freeze said. “AI learns from data sets, and right now, there’s a lot of interesting dialogue about how AI informed by personal data could aid in coronavirus response, like to more quickly identify a potential vaccine.”
“The MIT Technology Review recently had an article about how coronavirus is forcing a trade-off between privacy and public health,” he noted. “Again, what’s interesting is nearly one-third of our survey respondents said they are now more willing to share anonymized personal information if it will make their community safer or healthier.”
“So for right now at least, many consumers are ready to make that trade-off,” said Freeze.
Robots and AI approved for healthcare
The ongoing shortage of healthcare personnel and the need to protect them and patients has accelerated rollouts of service robots in hospitals.
About a third of people surveyed said they were more comfortable with AI and automated tools such as the Center for Disease Control’s chatbot to determine whether they need care, while one-quarter said they felt less comfortable.
In addition, 35% of respondents said they are now more comfortable with AI-driven screenings to determine whether someone can enter a healthcare facility, while 27% said they felt less comfortable.
While it isn’t yet clear whether attitudes will eventually revert to pre-pandemic levels, change is inevitable, said Freeze.
“I think attitudes will shift further towards comfort and acceptance as the presence of AI continues to grow in everyday life,” he said. “Regarding the pandemic, I don’t think we’ll hit a point where it feels like a switch has been flipped and the crisis is ‘past’ — at least in the next couple years — but when the novel coronavirus is no longer a threat, maybe when we have a working, widely distributed vaccine. I don’t think the comfort that has been built with AI and robots will go away.”
Delivery robots pull ahead of driverless cars in survey
More people are comfortable with robotic deliveries than with the idea of riding in a self-driving car, found the Interactions survey. If they knew an online order would be delivered by a robot, 73% of respondents would still place the order. Forty percent said they are more comfortable with delivery robots, but 23% said they are less comfortable.
A larger-than-expected number of respondents — 25% — said they would not order items online if they knew they would be delivered by a robot.
Attitudes toward autonomous vehicles have not changed, even as major automakers and technology companies have invested billions of dollars in research and development in the past few years. Fifty-five percent of survey respondents said their feelings have not changed, and 26% said they are less comfortable with the concept. Only 18% said they are more comfortable with the idea of riding in a self-driving car.
“Self-driving cars have a ways to go before they enter the mainstream,” observed Freeze. “Ultimately, people will have to be convinced that they’re safe, and there’s a lot of trust that needs to be built before that’s a popular opinion.”
Robots versus jobs?
Industry observers have pointed out that the current massive unemployment is the result of temporary business shutdowns rather than automation, which was steadily increasing during the previous period of low unemployment. At the same time, other research has found evidence of workforce changes.
Prof. Daron Acemoglu, an economist at the Massachusetts Institute of Technology, recently studied more than 55,000 French manufacturers, of which 598 purchased robots between 2010 and 2015. He stated that firms that adopted automation more quickly also hired more people than those that were slower, even as one robot replaced 3.5 workers. Acemoglu has also suggested that automation has displaced low-skill workers and exacerbated income inequality.
“I think the fear of job displacement plays a significant role in any lingering discomfort with AI and automation,” said Freeze. “I host a podcast about AI, “The ConversAItion,” and in our very first episode, we brought on recognized AI expert Rob Atkinson to explore this topic.”
Still, the global demand for robots, particularly cleaning and disinfection robots and mobile platforms, likely will continue to grow.
Designers must address skeptics
As robots and AI become more common in customer service, will attitudes become more positive?
“I’m not surprised that a quarter of respondents expressed discomfort with AI in their daily lives,” said Freeze. “I think there’s still a lot of skepticism around AI and automation in general — many people don’t have a clear understanding of what it is or how it can be used to augment human intelligence, rather than replace it.”
“As more robots appear in our daily lives, I expect people will become increasingly comfortable with their presence,” he added. “As with all technological change, it takes time for innovations to be normalized.”
Service robotics and AI companies must account for the distrust of a significant portion of the public, Freeze said.
“For developers designing hardware and software, I think it’s important to acknowledge that the coronavirus era is changing most consumers’ feelings about AI, meaning people are questioning and especially mindful of the way AI appears in their everyday life right now,” he said. “With that in mind, I believe it’s more important than ever to design for transparency. Developers and designers must do what they can in the creation process to make it extra clear to the end user why or how an AI system should be used. Otherwise they risk confusion, which can trigger unease and cause discomfort in consumers.”
“For example, at Interactions, we build AI-powered intelligent virtual assistants, and we believe a virtual assistant should never attempt to mislead the consumer into thinking it’s an actual person,” he noted. “We recommend having the assistant start conversations with a quick disclaimer explaining, ‘Hi, I am an intelligent virtual assistant, but you can speak to me just like you would a human. How can I help you today?’”
“For physical robots, too, it’s important that the technology is built with the end user in mind,” said Freeze. “In the design process, for example, developers can ask questions like, ‘How will the shape and appearance of this robot help inform onlookers about its intended function?’ The less surprised consumers are about a robot’s activity, the better.”
Mureli Muniandy says
Humans will willingly live and work alongside AI enabled robotic applications as long as ASIMOV’s fundamental ” 3 Laws of Robotics” can be practically demonstrated. For all the intelligence we have built into software and safety into hardware, we are still lagging behind when it comes to fulfilling the 3 laws.
William K. says
Why would I ever trust AI for service, when my present experience has been that it is quite unsatisfactory by a large amount. Meijer store customer service was a prime example. I went to the website to determine if the secret date code on a product indicated that it had expired, and what I got was an effort to create a charge account. Even a very unskilled human would not have made such a dumb mistake. Additional attempts have not done any better. So at least that type of customer interface AI is rather worthless.
The universal flaw with most AI is the lack of any ability to deal with any exceptions from the script.