New forms of "empathetic computing" are helping human users feel more comfortable in opening up to a program.
Why it matters: Our mental health has taken a major hit during the COVID-19 pandemic, while social distancing means it's harder to meet in person with therapists. That has opened a space for emotionally attuned machines to help us.
By the numbers: A survey of thousands of employees and executives released earlier this month by Oracle and the HR research and advisory firm Workplace Intelligence found 68% of people reported they would prefer to talk to a robot over their manager about stress and anxiety at work.
- 80% said they were open to having a robot as a counselor or therapist.
- "Workers said that robots can support their mental health better than humans because they can provide a judgment-free zone," says Dan Schawbel, managing partner of Workplace Intelligence.
Of note: In April, the FDA suspended many of its rules for digital therapeutic devices for psychiatric disorders because of the pandemic, which led to an increase in doctors prescribing new forms of digital therapy.
What's happening: Last month Maslo.ai, a startup that works on AI-driven empathetic computing, released new open-sourced toolkits that allow developers to build apps to address mental health.
- Maslo uses signal processing techniques to "read" voice, text and even the body language of a human user to identify a baseline level for mental health.
- "We can extract the linguistic aspects of what a person is saying, but we can also look at acoustic elements as well — volume, loudness, intonation," says Ross Ingram, Maslo's CEO.
- That data can help Maslo's partners build better executive performance coaches, for example, or even help with online dating.
Between the lines: A study by researchers at Maslo and the University of British Columbia found that while robots overall were seen by humans as having significantly weaker qualities of mind compared to people, younger users were much more likely than older generations to see machines as potentially capable of processing emotions.
Reality check: AI cannot yet feel or process emotions in any real way, which limits the effectiveness of therapeutic robots.
Go deeper: Better ways to use — and measure — AI in medicine
