Who Owns the Future of Suicide Risk Prediction?


On my plane rides to and from a recent meeting about big data in suicide prevention, I finally read Jaron Lanier’s 2013 book Who Owns the Future?  If you’ve read it -or read much about it – you can skip the rest of this paragraph.  Lanier is a legendary computer scientist often cited as the creator of virtual reality.  The first half of his book (my eastbound reading) is a Jeremiad about the modern information economy de-valuing human work and hollowing out the middle class.  Jobs are disappearing, income inequality is increasing, and personal information is devoured and monetized by big tech companies.  While Lanier gives a free pass to scientific users of big data, I think some of his criticisms still apply to our work.  The second half of the book (my westbound reading) proposes a solution.  It’s a new economic model rather than a new regulatory structure.  Lanier argues that those who create useful information should be paid by those who profit from it.  If you discover a secret shortcut around rush-hour traffic, then Google Maps may route other people to follow you.  Your shortcut might be ruined tomorrow, but your bank account would show a micro-payment from Google for “selling” your discovery to other drivers.  If you are an expert discoverer of efficient driving routes (like my wife), then Google might pay you over and over for the valuable driving data you create.    

Of course, individuals can have different preferences about sharing information.  In Lanier’s scheme, each person could own their data and set their own price.  Some people could choose to donate data for public good – not expecting any compensation.  That would be their perfect right.  At the other extreme, some people could set a price so high that no one would be willing to pay.  That would also be their perfect right.

In contrast to many big data skeptics, Lanier does not argue for requiring “opt-in” consent for data use.  In fact, he argues that the “Terms and Conditions” and “End User License Agreements” we click through so frequently are the antithesis of informed consent.  If we are asked for permission more and more often, we will just pay less and less attention to each request.  He proposes fair compensation as a just and more respectful alternative.

Lanier does admit that micro-payments for data use could be more of a thought experiment than a practical solution.  Even if our thousands of daily decisions create value for someone else, we don’t have time to send out all of those tiny invoices.  But Lanier thinks the internet could manage all that accounting for us – if we decided that it should.  And he certainly knows more than I do about what the internet can do.

Even if Lanier’s proposal is just a thought experiment, it did get me thinking.  His central point is that most so-called artificial intelligence is simply re-packaging of decisions by large numbers of old-fashioned intelligent humans.  People discover and create.  Machines just copy and aggregate.  Our current economic rules reward the copying and aggregating rather than the original discovery and creation.  If humans created the original value, shouldn’t they share the eventual profits? 

At first, I couldn’t see how Lanier’s proposed solution would apply to our use of health records data for suicide prevention or other public health research.  We don’t sell or profit from health information.  Answering scientific or public health questions won’t destroy middle-class jobs in health care.  But (one of the few good things about a 5-hour plane ride), I eventually saw how his proposal applies.  Even if no money changes hands when we use health data, we could do a better job acknowledging patients as the true creators or authors of health knowledge.  My friends at the Depression and Bipolar Support Alliance helped me understand that people who live with mood disorders are “experts by experience.” That expertise often goes unrecognized.

If we are using health records data to predict or understand suicidal behavior, then data from every member of our health systems could be useful.  But data from people who attempted suicide or died by suicide would be especially valuable.  We’d also highly value data from people at high risk who did not attempt or die by suicide.  Living with suicidal thoughts is hard work. 

Lanier’s proposed economic model reminds us who created the expertise that lives in health data.  Since it’s close to the 4th of July, I can’t resist a reference to “of the people, by the people, and for the people.” 

Greg Simon