Follow Datanami:
May 11, 2020

‘COVID Phobia’ Shifts AI Sentiment

via Shutterstock

COVID-19 has significantly shifted consumer sentiment towards automation technologies, including robot deliveries and AI-based chatbots used in telemedicine applications. However, a new AI sentiment report also finds the novel coronavirus has failed to move the sentiment needle when it comes to use cases like riding in self-driving cars.

The consumer sentiment survey released by conversational AI developer Interactions also confirms what others have reported, at least tangentially: Hospitals visits have plummeted during the pandemic, with critical ill patients dying at home out of fear of contracting COVID-19 while hospitalized. The healthcare news web site STAT reported in April that a survey of nine major hospitals found that the number of severe heart attacks treated at U.S. hospitals dropped nearly 40 percent since mid-March.

STAT labeled the sentiment shift “COVID phobia.”

That fear factor appears to be driving a shift in how consumers view AI and automation technologies in the era of social distancing. The Interaction survey of about 1,000 U.S. residents was conducted between April 18-19. At that point, there were more than 40,000 confirmed coronavirus-related deaths in the U.S.

While one-third of respondents said they were generally more comfortable with robots assisting doctors and nurses, that attitude shift was limited to narrow applications like use of chatbots and AI-based screening to determine if medical care or hospitalization was needed.

While not specifically mentioned in the AI sentiment survey, medical technologists note that AI technologies have also proven useful for tasks such as diagnostic imaging where human “confirmation bias” remains an issue.

Along with healthcare and telemedicine, the survey identified shifting attitudes toward greater use of automation in areas ranging from delivery robots to automating grocery store logistics for tracking inventory and cleaning. Indeed, automating grocery store functions rated highest in terms of shifting AI sentiment at 57 percent. One-third of survey respondents said they would be “comfortable with the idea of robots in grocery stores.”

Evolving perceptions about formerly routine but now stressful chores like grocery shopping were likely influenced by recent supercomputer simulations of how pathogens like COVID-19 spread in public spaces like grocery stores.

While consumer attitudes about robot deliveries is shifting, 55 percent of those polled by Interactions said the pandemic has not changed their view of riding in self-driving cars: They’d rather walk.

The survey also sought to gauge shifting sentiment about AI and privacy, especially the trade-off between public health and personal privacy. That debate is being played out with the rollout of a contact tracing framework that would be used to alert persons interacting with infected persons.

Asian countries took an aggressive and sometimes heavy-handed approach to collecting personal data in attempting to stop the spread of COVID-19. Taiwan, for example, largely succeeded. The privacy debate in the U.S. and Europe has focused on where personal data is stored and the need for voluntary, opt-in/out choices for contact tracing. A voluntary system with personal data stored on devices rather than on government or other central databases is seen as most effective from both a privacy and public health standpoint.

The AI sentiment survey confirms deep divisions about data privacy, with indications that public health concerns may be shifting attitudes. Interactions reported that 31 percent of respondents are willing to share anonymized personal data if they are used to “make their community safer or healthier.”

Overall, 56 percent said they are “generally willing” to share personal data while others remain reticent.

Recent items:

Detecting Consumer Signals in the 90% Economy

Intel’s Sentiment Algorithm Needs Less Training Data