The Second (Android) Coming of Philip K. Dick

There's a great science fiction story called "The Electric Ant." In it, an injured man is taken to the hospital, where operating surgeons discover that, unbeknownst to him, he's an android. After hearing the news, the man realizes his reality is dictated by reels of punched tape in his chest. He begins to tamper with the reels. After punching a new hole, a flock of ducks flies through his room. When he cuts the tape entirely, the world disappears.

The story is by the late Philip K. Dick, the legendary flipped-out weirdo and science fiction author whose books have been endlessly adapted into films. Everyone knows Blade Runner, based on his novel Do Androids Dream of Electric Sheep?

What most people don't know is that Philip K. Dick has been reincarnated as an android.

The story begins in 2004, when a team of roboticists and intelligent systems designers built a functional robotic likeness of Philip K. Dick. It looked like Dick, dressed like Dick – down to the gauche floral button-ups – and could hold forth on a wide range of subjects pertaining to his work. A low-res YouTube clip of the android in its prime shows it casually propped up on a brown divan, servomotors quietly abuzz, calmly informing its interlocutors that he was not a writer, but a "fictionalizing philosopher."

It's all very cute, except Philip K. Dick – the writer, not the simulation – was consumed, for much of his career, with the idea that a sophisticated-enough android, one virtually indistinguishable from a person, would present us with complex moral and metaphysical questions about the nature of identity. Dick's paranoid-android oeuvre is broad and impressive. He penned futures rich in personable machine intelligences, often tinged in moral ambiguity.

Dick was also famously unhinged – he suffered a psychic break in 1974 after which he started being visited by a "transcendentally rational mind" in the form of a sentient pink beam – and the question of personhood in the machine age possessed him. He didn't see a huge leap, or even much of a distinction, between this and the bigger question: what is reality, anyway? What if you woke up one morning to realize your reality was a construction, your memories implanted? Would you suddenly be less human? Would the world be less real? In a 1978 essay, Dick wrote:

Fake realities will create fake humans. Or, fake humans will generate fake realities and then sell them to other humans, turning them, eventually, into forgeries of themselves. So we wind up with fake humans inventing fake realities and then peddling them to other fake humans.

In 2006, Philip K. Dick's android was traveling to Mountain View, California, to meet with Google employees, when it vanished. It was apparently lost by an airline, but speculation abounded. Did the droid go rogue, or was it stolen? Was it shivering in a pile of garbage somewhere, having a hellish time of consciousness, desperately seeking a sympathetic human to abet its escape?

As a tale of life-imitating-science-fiction, the story could have ended here, except that like the man, the android has been reincarnated. This new android is being referred to as "New Phil," its vanished predecessor, "Old Phil."

So, just to recap: a man who spends a career writing metaphysical novels about robot identity dies. Twenty years later, an android is made in his image, effectively bringing him back to life. That android disappears under mysterious circumstances. A new one is built, with a completely different machine intelligence, which brings us three degrees of separation from the original. Dr. David Hanson, the roboticist responsible for the PKD android's mechanized, lifelike face, says that despite the marked improvements in the second version, he's still "chasing the essence of the real PKD," adding that "maybe the Android-Phil unit #3 will channel more of his true essence, when we get around to building it."

Much of the source code for the original android is available for free online, which means that one can conceivably conjure up a personal Philip K. Dick from off-the-shelf parts. It's a story Dick himself could have written, perhaps with some future model (New New New Phil) becoming self-aware. Surrounded by dozens of alternate versions of himself, he slowly becomes convinced that he is the real, original Phil. Did you know the "K" stands for "Kindred"?

It's hard to guess what the writer, who died in 1982, would have thought of all his synthetic descendants. The lead programmer on the original PKD android, Andrew Olney, now the associate director of the Institute for Intelligent Systems at the University of Memphis, speculates that, "he would have appreciated the reification on a certain level…to the extent that it fell short of the androids he wrote about, I think he would have been relieved. After all it is not the basic idea of an android that is the most troubling, but rather the idea that an android can be more real, or more human, than a human."

The androids in Dick's stories are clearly sentient beings, in the sense that they feel, fear death, and experience transcendence. But are they people – do they have selves? Could they have "a soul?" And can we feel empathy for them? We can't answer these questions until we define the difference between existence and consciousness, or what it means to be a person.

For our human purposes, there might not be a practical difference. That is, after all, the essence of the Turing test: that artificial intelligence can be considered achieved the moment a computer fools a person into believing it's a human being. It doesn't take much for us to relate to the inanimate; sometimes all you need to do is stick a pair of googly eyes on a fire hydrant and suddenly everyone is like, "aw, hey little guy!" (the phenomenon is called pareidolia, incidentally).

And while I'll never get to chat with the real Philip K. Dick, I doubt I'm the only head who'd kill to have an hour-long conversation with his double. Or triple, rather. It's close enough. Maybe, even, it's closer: the android is a person that PKD himself could only speculate about being. More human than human.

In designing the artificial intelligence that compelled the original robo-Dick towards discourse, Olney had to take a block of conversational Dick (mostly interviews and letters) and use it to build a massive semantic space, a vector space that mapped words to vector representations. The end goal was to create a giant memory of conversational stimulus and response pairs, so that the android could hear a question and respond with answers culled from a database of Dick-isms.

For every spoken input, the android had about twenty responses at the ready. When posed a question with more than one possible answer – which is to say, any interesting question – it would respond with the centroid, or most prevalent, answer. Of course, that's not how people work. We don't average our responses from a field of potentials; instead we're always developing ideas, adapting to people around us and our own changing sensibilities. Even our sense of self varies from year to year, from place to place.

Philip K. Dick was uncannily adapt at perceiving precisely this kind of human mutability. Identity is dependent on a list of subjective variables, all of which can be shifted (or shattered) through a forced change in perspective: tragedy, drugs, education, pink beams of light, or discovering tape-reels spinning in your chest, for example.

The line that divides the individual from the world, and thus defines the self, is always in flux. And in a technological milieu, the digital amnion we now call home, these boundaries are even more porous. I'm not an android – at least I think not – but I am a cyborg. I am increasingly dependent on a series of external motherboards and hard-drives to process and store my memories, experiences, my social life. As I perpetually curate my identity online, I'm simultaneously being perceived and unpacked by an invisible audience (which, by the way, hi!).

The more we define who we are using technology, the more technology begins to dictate our sense of self; the crux of Dick's canon isn't literally about being an android – or even making one – but rather the inevitability of an intermixed self and machine, in whatever manner our generation comes to invent. With that inevitability comes two things: a sudden quantifiable kind of identity, controlled and subject to malfunction. And an increased ontological chaos, a wildly propagating fractal subjectivity, where the question "what is real?" depends less on where you are than on where you point your browser.

To make the world disappear, Dick's android only had to cut the tape in his chest, but we may already be halfway there.

• Earth, a new novel by PK Dick
• That Fear of Machines Building Machines
• A "Social Roboticist" Wants You To Hug Your Robot