“Everyone must leave something behind when he dies, my grandfather said. A child or a book or a painting or a house or a wall built or a pair of shoes made. Or a garden planted. Something your hand touched some way so your soul has somewhere to go when you die, and when people look at that tree or that flower you planted, you’re there.”   – Ray Bradbury, Fahrenheit 451

In a 2013 episode of the British science fiction anthology series Black Mirror, a woman whose partner dies suddenly in a car accident enlists the aid of a new service to help her overcome her grief.  The service ingests her partner’s digital footprint – his texts, social media posts, video and audio – and creates an artificial intelligence version of him that mimics his voice and thoughts.  It works, for a while, but soon creates more problems than she ever thought possible.

Billions of people have now created digital lives on social media platforms.  We spend hours cultivating these presences, interacting with others who are doing the same thing, but give little thought to what will happen to our digital selves when we die.  A new wave of companies aim to commoditize a digital eternity for each of us – a way to keep us alive, even if just a pale digital shadow of ourselves, in the hopes of comforting our loved ones.

This situation began to gnaw at Rebecca Blum, a senior strategist at San Francisco design firm frog, who took time away from her usual work to study the intersection of death and tech – or necrotech, as she calls it – and presented her thought-provoking findings on Saturday at SXSW in Austin.

The genesis for Blum’s exploration came from a friend who signed up for a dating app called Hinge, which mines your own social networks to find potential mates.  The friend got connected to a man who she has now been dating for a while – but their mutual Facebook friend Heather, whose network led Hinge’s algorithm to connect them, died in 2015.  The couple find this connection inspiring and saddening in equal measure.

“Essentially, they were introduced to each other by Heather’s ghost,” Blum said.

Facebook acquires an average of 50,000 new users per day, but 10,000 other Facebook users die each day, according to Blum.  Those departed users leave behind countless text posts, photos and videos – and similar patterns are repeated across Twitter, YouTube and other platforms.  And this doesn’t even account for e-mail, where companies like Google have amassed vast troves of archived, searchable data tied to specific individuals.  The BBC called this “an unstoppable digital graveyard”.

These tech companies have been so focused on user growth and monetization that they haven’t done much thinking about the issue, either – at least publicly.  Blum  said Twitter recommends appointing someone to delete your account upon your death, Yahoo will clear out your accounts if someone sends them a physical copy of your death certificate, and Google has an inactive account policy (which Blum calls the “We Think You’re Dead policy”). Facebook has made the biggest strides with their legacy contact functionality; you assign a specific person or persons to have access to your account if you die, and they can convert it to a legacy account where the word “Remembering” is added to your profile.

In addition to social platforms, startups have begun to disrupt this space:

• Japanese app Suma Tomb gives users the ability to create an augmented reality video of themselves, which after their death can pop up and play for anyone who visits their grave (or a virtual version of it) and enables the app.
• Portugese startup ETER9 (a mashup of ‘eternity’ and ‘cloud nine’) allows users to create a virtual being that, via machine learning, processes their posts and comments they leave on the platform to eventually reach autonomy. After two years in beta, ETER9 has more than 40,000 users and plans to continue expansion.

Others are experimenting with post-death AI as well.  San Francisco software engineer Eugenia Kuyda, grieving after the loss of her friend Roman Mazurenko in an accident in 2015, created a “Romanbot” – an interactive textbot that mined all the text messages and other correspondence Roman had sent her over the many years of their friendship, and was able to respond to texts she sent it with a tone similar to Roman’s when he was alive.  She often texted it when she felt sad or thought about him, and soon other friends of Roman’s began to do the same.  Yet, other friends were disturbed by the prospect and refused to interact with the bot, according to a Verge article about the bot experiment.

“This is all very bad,” friend Vasily Esmanov said, mentioning that he felt Kuyda hadn’t learned from the Black Mirror episode. “Unfortunately you rushed and everything came out half-baked. The execution — it’s some type of joke. … Roman needs [a memorial], but not this kind.”

Blum asked the SXSW audience, whom she praised for being willing to get up early on a sunny Saturday morning to come hear a talk about death, how they felt about such efforts.  A show of hands showed that many felt they were creepy and crossed a line, yet a few said they were comforting, and even were interested in creating a bot of themselves.

While many of us using social platforms today may recoil at the prospect of an AI version of ourselves living on and interacting with those we leave behind, Blum thinks that future generations – like current toddlers who already know how to take a selfie – may be so immersed in the world of tech that such ideas won’t necessarily seem strange at all.  Present-day futurists have predicted this for a while; Ray Kurzweil said that one day man will consider the idea of having one physical body and no “backup of our mind file” as a primitive concept at best.

Japanese roboticist Masahiro Mori developed the idea of the “uncanny valley,” a place where robots that continue to be more and more humanlike will be more endearing to humans – but yet the likeness is a little too perfect, and slightly “off”.  Blum is concerned that the technology behind AI and digital interaction is creating what she terms “the uncanny death valley.”

“It’s getting harder to separate the ideas of physical self and virtual self,” Blum said.  “We’re not very good at distinguishing between what’s living and what’s dead online.  … How can we make this interaction more thoughtful and less creepy?”

Here are her takeaway ideas to address the problem:

• Be more intentional. We make wills, we set aside time to mourn, we build monuments or memorials – but when we go online we don’t put such careful actions around it. We should develop a set of standards or guidelines about how our information is to be used – Blum compared it to the organ-donor designation declared on drivers’ licenses – so that our loved ones and tech platforms alike know what needs to happen to our digital presence when we die.
• Allow for distinct transitions and spaces. Blum noted that many faith traditions embrace the idea of death as an “other” – heaven for Christians, Hades for ancient Greeks, and so on. Even in our earthly life, cemeteries often have elaborate gates to provide a signifier that you’re moving from our regular everyday world into a different space, marking the way of death.  Yet, digital applications like Facebook often don’t make any separation at all, and when they do it’s usually marked by barely noticeable differences. In our digital spaces, those divisions need to be made more clear. Blum gave an example of a Holocaust survivor, Eva Schloss, whose story is being preserved through an immersive digital project.  She noted that, while giving her a lasting digital presence that will outlive her, it’s done with appropriate separation since it will live as part of a museum exhibit.

• Think about ways to allow mourning to slowly decay: In the Jewish tradition of shiva, loved ones mourn a death intensively for seven days. Then, a year later, you gather those people together and mourn again. Blum cited this as another example of giving mourning specific spaces; “you can’t be in that headspace all the time,” as she put it.  With post-death AI situations like Romanbot, where the digital loved one is everpresent, it may inhibit the ability of loved ones to mourn effectively and then move on.  But, if done very intentionally, it could allow you to slowly decay your relationship with that person in a less intense way.
• Find opportunities for tangible traces: What if there are more visible and digestible ways to leave behind a digital legacy? For example, Blum cited Google’s My Activity record that shows what you read and did online every day.  What if there was a more curated way to do this – and one you can opt in for, as a way to clearly leave behind a specific version of you?

Social psychologist Clay Routledge, in a 2014 piece on how we think about meaning in our lives, came to the conclusion – which Blum cited in her presentation – that a sense of purpose drives us to live longer, and think less about death – yet still hope for something that outlasts us.

“People know their lives are brief,” Routledge said.  “So we endeavor to be part of something that transcends biological existence.”

Whether we believe in an afterlife or not, we still want to be remembered here. There’s no reason we can’t leave behind meaningful digital memories, too.