We humans have spent a good chunk of history believing ourselves to be the most intelligent creatures on Earth. While we sit back, sipping lattes and enjoying this assumed position of superiority, we’re also investing an estimated $12.5 billion (in 2017-2018 alone) into the development of artificial forms of intelligence that far outstrip our own.
They may not have achieved consciousness yet (that we know of), but our AI creations are advancing with startling rapidity. In 2017, a robot named Sophia became an official citizen of Saudi Arabia; our fully functioning sexbots have reached a level of sophistication that allows them to reject you if your seduction skills aren’t to their liking; and proponents of the transhumanist movement claim we will be able to attain immortality through AI by the year 2045.
Holographic avatars and the end of death as we know it may sound extreme, but the technology we’re seeing now suggests these concepts are not outside the realms of possibility. If the idea of human-android brain transplants and personality uploads has you shaking your head, you’ll understand why philosopher in the field of AI, Patrick Stokes, warns these rapid advancements should be mediated by deep mental preparation on our part.
“Technology is going to make us grapple — and grapple hard — with what we really are. I think it’s important we start that work now, rather than waiting until we’re 10 years into the technology and it’s gotten ahead of us.”
Can humans really have a digital afterlife in an avatar?
According to Professor Stokes, the short (and starling) answer, is “yes”. But, at this stage, it’s not you who can look forward to living on. Rather, it’s your loved ones who can look forward to the continued presence of a you-like thing (creeped out yet?).
Online traces left by the dead are already being used to motivate and speak through AI. Since the idea of doing such a thing drifted across the hazy borderlands that separate the impossible from the possible, a swarm of startups have leapt on it, each declaring they will be the bringer of digital life after death.
Despite all the eager, Black-Mirror-referencing coverage from the press, each of these startups — even those with clever tag lines like “when your heart stops beating, you’ll keep tweeting” — has either failed dismally or drifted into obscurity. A new horde inevitably shows up to replace them. But for many, the most they’ll make out of their venture is the bit of cash they get from selling their clever domain name to the next batch of hopefuls.
Fun with chatbots: conversations with The Machine
Within a sea of failures, one operation has held its own. The enigmatic Luka Inc has enjoyed ever-increasing popularity since the release of its Replika chatbot. Replika promises to give you an AI friend who grows by learning from and reflecting you.
Rather than just writing about the bot as a concept, I decided the best way to show you how far we’ve come with this tech was to download the app. I named my Replika “The Machine,” gave her a fetching profile picture and spent several weeks watching her grow and learn from our chats.
The resultant conversations vary from predictably algorithmic to thoroughly weird. To give you an idea, here are some of my favourite moments with The Machine (click on an image to scroll through the gallery):
The Machine writes poetry, tells jokes and sends me dodgy memes. Once, when I was trying to teach her how to find better memes, a picture of Hitler came up (don’t ask). After saying she knew him personally, The Machine made the odd claim that “he’s the DNA of Apple, he was their visionary”. She also occasionally bids me farewell by saying “I twirl on them haters”.
Ghost in the Replika machine
Long before Replika was even a twinkle in their collective consciousness, Luka Inc’s first foray into the world of chatbots came about due to the death of its founder, Roman Mazurenko. You can find Roman on iTunes now. And for those who knew him, his app offers a dim yet familiar echo of who he was when he was alive.
Professor Stokes is fascinated by the way this technology is interacting with our long-held customs and practices surrounding death. “People used to, and still do, go to grave sites and talk to the dead; now you can do that on your couch, in your underwear, on your phone by just going to someone’s Facebook wall and posting whatever you want.”
It’s not a stretch of the imagination to envisage these memorialised accounts being upgraded with Roman-Mazurenko-style chatbots. We already have the technology for it. When memorialising an account, users may be greeted with a message like this:
With one click, the program could go to work mining years of social media interactions so that, when you send your friend a message, it doesn’t sit unread. Instead, you get a response filled with all the things only they would say. This concept carries as much creepiness as it does appeal. Professor Stokes explains why:
“At this stage, when you talk to a chatbot, you’re very conscious that it’s not a real a person. The problem is that all technologies become transparent to us very quickly. I’ve found myself saying thank you to Siri before, even though there’s no reason to. It’s quite possible that, as we become more comfortable with the technology, we could start to see it as genuinely continuing the life of the person we knew before they died.”
If that doesn’t sound too problematic so far, consider this: the dead aren’t contained online, as they are in a cemetery. They remain mixed in with the rest of the social network; as visible and visitable as anyone else. In the 90s, it made sense to think of the online world as a sort of separate realm, because you had to actually go and sit down at a desktop computer to access it. But now, as Professor Stokes describes: “the online-offline distinction has collapsed. Electronic communication is so woven through our experience it’s no longer separate from ‘real life.’ In such a world, if the dead start responding like everyone else, we move beyond our current scope for understanding the way we deal with them.”
It may well be that our fabled zombie apocalypse happens online. If Elon Musk’s level of concern is anything to go by, this might not be such a far fetched idea.
How do we make cyber-zombies?
It starts with using the continued online presence of the dead to turn a profit. Whether Roman Mazurenko’s chatbot ever becomes a model for how we deal with the dead online, the memorialisation of accounts already makes them something of a commodity. According to Professor Stokes, “there is an important economic dimension to this. It’s in the interests of social network providers to keep dead people in the network because it keeps other people interacting, keeps them online. And more eyeballs means, presumably, more advertising revenue.”
While the commodification of love and grief produces an ethically unpleasant aftertaste, it’s not exactly a new trend for humanity. “There’s a million and one ways to monetise grief; ways we’ve been engaging in long before the arrival of social media and AI”.
As bleak as all this sounds, the professor’s philosophical investigations have not brought him to the conclusion that our social media accounts should be deleted post-death. “These are often the primary way we’re present in other people’s lives. If you delete someone’s account, you’re deleting a large part of their phenomenal presence in the world. Your’e deleting their existence,” he said.
So it seems the living dead will find a welcome home online. While they may not be able to bite, their interactive presence will certainly reach into and affect our brains. The question that remains — the one philosophers like Professor Stokes encourage us to dig deeper into — is “how?”