A new report today highlights anxiety that U.S. consumers have over the possibility of artificial intelligence going too far. Despite advances in robotics and self-driving cars, a majority of Americans reported mixed feelings about the benefits of AI, according to a study by Elicit.
In its 14-page report titled “Artificial Intelligence and the Very Real, Real-World Anxiety it Causes,” the research consultancy found that news reports over intelligent personal assistants that have “gone rogue,” as well as Facebook’s misuse of personal data, have created distrust among consumers.
Approximately seven out of 10 (73%) think that some companies will go too far with AI, and another six out of 10 “are concerned with how companies will use AI and the information they have to engage with them.”
“AI can be a transformative way for companies to interact with their customers and deliver better experiences,” said Elicit CEO Mason Thelen in a press release on the study results. “But an important perspective was missing in this conversation – how consumers feel about the use of AI in their everyday lives and by companies they engage with.”
The survey was conducted in Spring 2018 with a nationally representative sample of 697 adults, Elicit said. Additional findings from the study include:
- Fifty-nine percent said that AI “has the ability to be good, but has some inherent risks associated with it.” Meanwhile, 18% agreed that AI “will pose an existential threat to humanity,” while 15% said, “It is a benevolent force.”
- While 38% agreed that “companies should use AI to make my experiences better” (39% were neutral, and 23% disagreed), a lot more people were concerned (64%) about how companies would use AI and the information they have about them to engage with customers.
- Concerns about whether AI will be “better than humans” are not as strong. Sixty-four percent said that “artificial intelligence will never be as good as human interaction,” while only 14% disagreed with that statement.
Benefits of AI differ depending on human input
The study also looked at how people responded to scenarios based on whether human interaction was needed by AI elements. For example, in an interaction between Alexa and a smart refrigerator, respondents became more nervous with the scenario “Your smart fridge tells Alexa that you are low on milk, butter, and eggs; Alexa orders groceries with free same-day shipping and notifies you” (63% were nervous, with 37% “happy”) than one where the human notices that they are low on milk, butter, and eggs, and then notifying Alexa to place the order (74% happy, 26% nervous).
Similar nervousness responses came from scenarios involving autonomous cars that drive automatically and make decisions to minimize damage during unexpected dangerous situations. Respondents were also wary of a hypothetical dating site AI that “predicts a soul mate based on a series of questions that you answer, knows that you frequent similar establishments in your neighborhood, and schedules a date at one of those restaurants.”
In addition, they expressed concern about a subscription clothing retailer that “captures what you wear on a daily basis by a camera over an extended time and predicts new items you will like and automatically ships them to your address.”
In the area of robotics, a large majority (82%) said they would be happy with a personal assistant robot that takes care of routine household chores such as vacuuming, dishes, and laundry, based on commands from a human. But they were less excited (51% to 49%) about a personal assistant robot that could identify new household tasks it could perform and just does them.
“Suggesting new chores is fine, but the personal robot likely needs to demonstrate its proficiency and seek approval before it proceeds,” the report stated.
The complete 14-page report is available for free here.