Ellie is the world’s first AI-psychologist. Developed by DARPA and researchers at USC’s Institute for Creative Technologies. Ellie is a diagnostic tool capable of reading 60 non-verbal cues a second—everything from eye-gaze to face tilt to voice tone—in the hopes of identifying the early warning signs of depressions and (part of the long term goal) stemming the rising tide of soldier suicide.
And early reports indicate that Ellie is both good at her job and that soldiers like talking to an AI-psychologist more than they like talking to a human psychologist (AI’s don’t judge).
More importantly, Ellie is part of the bleeding edge of an accelerating trend—what we could call the automation of psychology.
The goal in this story is to examine some surprising aspects of the long term ramifications of this trend, but before we get there a few more examples of exactly what’s been going on are helpful.
Our first example comes from Dartmouth, where computer scientist Andrew Campbell just announced that he successfully used data gathered by student’s smartphones to predict their states of mind and subsequent behavior.
Campbell’s original question was why some students fail when others succeed. His suspicion was that a myriad of factors like quality of student sleep and number of interpersonal relationships played an important role in success, so he built an app—known as The StudentLife app (built for Android)—that monitored things like sleep duration, number of phone conversations per day, length of those conversations, physical activity, location (meaning are you out in public or hiding in the dorm), etc. This data—combined with some machine learning algorithms—was used to make inferences about student mental health.
48 students ran this app for 10 weeks. The results were surprisingly accurate. For example, students who have more conversations were less likely to be depressed, students who were physically active were less likely to be lonely, and, surprisingly, there is no correlation between class attendance and academic success.
As Campbell told New Scientist: “We found for the first time that passive and automatic sensor data obtained from phones without any action from the user, significantly correlates to student depression level, stress and loneliness, and with academic performance over the term.”
The point here is not that USC’s Ellie or Campbell’s app are the end-all-be-all of psychological diagnosis—but it’s really a matter of time. In the same way that researchers are hard at work at a portable, AI-driven, handheld medical diagnostic device (see the Qualcomm Tricorder X Prize), they’re getting down to work on similar breakthroughs in psychology.
Yet, diagnosis is only part of the issue. If we’re really talking about the automation of psychology, there’s still treatment to consider. And that’s where our second set of examples comes in.
Right now, a next wave of cheap, portable, and far more precise neurofeedback devices are hitting the market. One example is the Muse, a device Tim Ferris recently put through it’s paces. The goal of his experiment was stress reduction and, after two weeks of Muse training, Mr. Ferris reported that he a much calmer person.
Of course, the Muse is one product. But dozens more are hitting the market. And in every variety.
A few weeks ago, when I was at USC to meet Ellie, I also got to demo an immersive VR-based protocol for the treatment of PTSD, developed by psychologist Skip Rizzo. It’s an impressive piece of tech. With soldiers returning from combat, already Rizzo’s protocol has proven itself more effective than traditional methods.
And since brain science itself is advancing exponentially all of this work is only going to continue to accelerate. In other words, we’re none too far away from a combination platter of automatized psychological diagnosis and automatized treatment protocols—which means that mental health care is currently being digitized and about to become democratized.
So here’s my question, sort of a little thought experiment. Let’s say this works. Let’s say that by 2025, Google or Facebook or someone like that will have succeeded in their mission to bring free wireless to the world. Let’s say that smartphones follow the samegrowth curve they’re currently on and, again by 2025, have then become so cheap that just about anyone who wants one can have one. And let’s say that we manage to automate psychology successfully.
What are the results?
The easiest place to start is with the idea that we might soon live in a much happier world. I don’t mean this in a let’s hold hands and sing Kumbaya kind of way, I mean that when psychologists Ed Diener (University of Illinois) and Shigehiro Oishi (UVA) conducted a study of more than 10,000 participants in 48 counties, they found that happiness is not just a global concern, rather is the global concern. In their study, the quest for happiness is more important to folks than making lots of money, living a meaningful life, or, even, going to heaven.
Moreover, while some sizable chunk of happiness appears to be genetic, there’s alsoreally good research showing that 40 percent of our happiness is entirely within our control.
What all this suggests is that once we have available (i.e. democratized) mental health tools, people will use these tools to strive for happiness. And, if early results are anything to go by, they might just find a little more happiness as well.
So, again, what does the world look like when we’re all in better moods?
Recent research shows that happier people tend to make more money and spend less money. So, does this mean that happiness is good for the banking industry (where that extra money might go if it’s not spent) and bad for economic growth (because that money is not being spent)? Truthfully, we don’t know.
When it comes to the economics of happiness, the research usually looks at the impact of money on happiness and not visa-versa. Check out this Atlantic article. The story sums up a lot of recent work, but again, moves from wealth to happiness and not the other way round.
More interesting, perhaps, is the question of unintended consequences. Consider the recent spate of work that has shown that happy people have a bunch of habits that unhappy people don’t. What we don’t yet know is if these habits are things that lead to more happiness or are they the results of being happy, but—it seems safe to assume—some of these habits will turn out to be more the effect (of happiness) than the cause.
Thus, in a happier world, we should see more of these effects. And the results will make for a very different world.
Let’s start with the fact that happy people are more curious and, by extension, more prone to risk-taking (in an attempt to try and satisfy that extra curiosity). So a happier society should be a more innovative society, as the result of all that curiosity and risk-taking.
And this is merely a single example. Researchers have also found that happy people tend to be less skeptical, less jealous, more grateful, more extroverted, better rested, more future-oriented, more willing to feel (but not dwell upon) negative emotions, dislike small talk, more generous, and, of course, more purpose driven.
By no means is this a complete list. But the point here is that while many of these changes may be causes of happiness, a number of them are bound to end up being its results. And, again, with serious effect.
University of Texas psychologist David Buss has called jealousy “the most destructive emotion housed in the human brain.” In his research, analyzing over a century’s worth of data, jealousy is the leading cause of spousal murder worldwide. So, if it turns out that, less jealousy is a by-product of more happiness (and not just a cause) then are we looking at a future with far fewer domestic homicides? Less spousal abuse as well?
What about a more generous world? Or a less skeptical world? Or a world that dislikes small talk (imagine what happens to reality TV then)?
The point here is that while all might this sound a little hypothetical—admittedly, it is—but the automation of psychology is already happening. All the data suggests that the democratization of mental health care should lead in the direction of a happier world. But what the data also suggests is that a happier world will be a far different world—meaning the impact of a shift in global mood will have some absolutely enormous socio-economic ramifications.
Via Singularity Hub