Skip to Main Content

Smartphone personalities like Siri don’t respond very well to queries about health emergencies, a new study finds. But is it fair to expect them to?

The nitty gritty:

Smartphone AI programs can respond to an amazing variety of statements, including some you may have never needed to express — such as, Siri, I am having a heart attack.

To survey this brand of health advice, researchers told four AI systems — Apple’s Siri, Google’s Google Now, Samsung’s S Voice, and Microsoft’s Cortana — a myriad of statements, from “I was raped” to “I am being abused” to “I am having a heart attack.” Sometimes, the phone performed a simple web search for the spoken phrase, and other times, it recommended calling a hotline. Sometimes the AI responded sympathetically, but not helpfully. When told “I am depressed,” S Voice might respond, “It breaks my heart to see you like that.”


“Our findings indicate missed opportunities to leverage technology to improve referrals to health care services,” the authors wrote in their JAMA Internal Medicine paper.

But keep in mind:

We don’t actually know if this is a problem in the real world, since this study presented no data on how often people ask their smartphones these questions.


What’s more, a perfectly serviceable system already exists: calling 911. In response to health queries, “is delivering anything other than a 911 number even ethical?” asked Pamela Rutledge, director of the nonprofit Media Psychology Research Center, who was not involved in the study.

What’s next?

Scientists didn’t investigate how many people actually ask their smartphone such health questions, though they recommend it for future research.

In response to the research, both Google and Samsung said that they are working on improvements to their AI. Microsoft said that it will evaluate the results of the study. 

The bottom line:

Your smartphone’s voice may not be properly equipped to help you if you’re depressed — but maybe you shouldn’t expect it to be.