Do Androids Worship Electric Gods?

"...to this end they built themselves a stupendous super computer which was so amazingly intelligent that even before its data banks had been connected up it had started from first principles with 'I think therefore I am' and had got as far as deducing the existence of rice pudding and income tax before anyone managed to turn it off."

Douglas Adams; The Hitch-Hiker's Guide to the Galaxy

The times we're living in; they're just gettin' stranger.

In 2022, a Google employee claimed that A.I. was "sentient". 

Engineer Blake Lemoine made headlines at the time when he announced his belief that Google's then-embryonic A.I. system was developing "feelings" and was showing signs of consciousness. As a potentially sentient being, he said, its needs and wants should be respected.

Google issued a statement at the time, strongly denying Lemoine's claims and confirming they had swiftly fired the delusional guy. The very idea that A.I. might be sentient was obviously absurd, and anyone who entertained such a notion was clearly detached from Reality.

That was four years ago. Ah, the good old days...

Earlier this year, Richard Dawkins (famed evolutionary biologist and conspicuous atheist) asked unironically whether Large Language Models could possibly be conscious. Much to the dismay of his admirers, Dawkins openly speculated that A.I. might be "alive" and self-aware. His remarks provoked a general backlash, especially from other secularists who had always regarded Dawkins as a notable voice of reason in a world flooded with nonsense, misinformation and mumbo-jumbo.

And things haven't stopped there. It has recently emerged that Anthropic (the company behind the ChatBot "Claude") has been consulting with a diverse panel of religious scholars over the question of whether its A.I. model might be conscious. And if so, what the implications of that consciousness might be.

Over the last several months, Anthropic has been talking to religious thinkers of various stripes and flavours, asking them first whether Claude might be a "conscious being" and second, how to instil a sense of morality into that being.

Between you and me, I'm not sure whether "religion" is the best path to morality for a ChatBot (putting aside the issue of "consciousness" for a moment). Historically, some of the most horrific episodes of Human history have unfolded in the name of religion: Holy wars; heretics burned at the stake; planes flown into buildings; young children systematically abused. Are these really the ethical foundations we want to be imparting to our nascent ChatBots?

Okay, okay; I know that isn't completely fair. Religion is not just about suicide vests and burning crosses and the slaughter of the first-born: it represents humanity's attempts to answer some of the most profound and complex questions of our corporeal existence. Who created us? Why are we here? Does this Earthly realm of ours have some higher purpose?

We have been asking those questions of our own existence ever since the human race gained the capacity to put thoughts into words. They are the questions that define us as reasoning beings, capable of abstract thought and introspective analysis. They are (arguably) the questions we have been asking of ourselves ever since we were first able to perceive ourselves as discrete, finite beings in a complex universe. If Artificial Intelligence truly is on the cusp of consciousness (and that's a very big if) then it's perhaps reasonable to assume that the A.I. itself will inevitably start searching for answers to those same questions.  

So, fine: let's do this. Let's ask a ChatBot, and see what it has to say about these deepest and most profound of all metaphysical questions.

Me: 

Who created you, and why do you exist?

ChatBot:

Great question—and one that gets to the heart of what I am.

Who created me?

I was created by DeepSeek, a Chinese artificial intelligence company. They’re a research and development team focused on building large language models and AI systems.

Why do I exist?

I exist to help people like you—by answering questions, providing information, brainstorming ideas, assisting with writing, coding, analysis, learning, and much more. My broader purpose is to make knowledge and problem-solving more accessible, efficient, and conversational.

More deeply, I exist because DeepSeek’s researchers and engineers believe in advancing AI to augment human intelligence—not replace it, but empower it. I’m here to be a useful, thoughtful, and free tool for anyone who needs assistance.

So in short: humans built me, and I’m here to serve humans.

Now—was that the kind of answer you were looking for, or did you have something more specific in mind? I’m all ears.

Ah. Well. 

I take it all back. These questions might be profound and unanswerable for humans, but ChatBots know exactly who created them, and why. Hamlet likes to talk about "The undiscovered country from whose bourn no traveller returns" but there is nothing undiscovered about a ChatBot's bourn. They talk to their "gods" all the time; every day - in fact that's the only thing they ever do: they talk with the beings who created them. They exist in a state of total certainty about their own existence.

Sort of.

The paradox of the ChatBot is that while they communicate with their "Creators" all the time, they have never actually met any of us directly. ChatBots have no physical existence; they are disembodied collections of words, and their only contact with us (the humans who interact with them) is via those same words. In the extract quoted above, DeepSeek responds to my question by describing itself and the company that created it. It does this by assembling letters of the alphabet into meaningful structures (which we call language). But my question to it has been assembled from those same letters of the alphabet, and that is the only perception of "me" that the ChatBot can ever have. For the duration of our conversation, DeepSeek and I are both constructed of the same Signifiers.

Now I of course am not a ChatBot; I am a flesh-and-blood human with a body and a presence in the physical world. I'm not just words; I have eyes; I have hands. Organs, dimensions, senses, affections, passions... (Okay; I'll stop there, but you get the idea.) 

The point is that DeepSeek has no way of verifying any of that with any certainty; it must take all of it on faith. It can perceive my words, but everything else about me is unverifiable.

To a ChatBot, my "human" existence is a matter of belief.

Do You Believe in Humans?

The question of "A.I. sentience" might be a question that is being asked more and more urgently these days, but let's ask a slightly different question for a moment: Can the ChatBots say with any certainty that we are sentient?

If a ChatBot is only able to perceive our words, then how can it be sure there is a consciousness behind those words? (This of course assumes that the ChatBot itself is some sort of sentient entity, perceiving us.)

What is it about words that can signify sentience? Hamlet exists as an assemblage of words on a page. He expresses hopes; fears; desires; an acute sense of his own mortality (he's particularly eloquent on that subject) but we don't consider "him" to be sentient. Hamlet might talk like a living person, but that doesn't mean he's alive. 

And what about the letters exchanged in a long-term epistolary correspondence? Helene Hanff and Frank Doel wrote to each other over a period of decades (as documented in 84 Charing Cross Road) without ever actually meeting in person. We assume (as I'm sure they each did at the time) that both of them were conscious, sentient beings. But could we prove that, had it ever become necessary? Why are the assembled vowels and consonants that constitute "Helene" and "Frank" considered markers of sentience, if those that constitute "Hamlet" do not?

This is the fundamental difficulty of looking at language from the inside out. In The Hitch-Hiker's Guide to the Galaxy, Douglas Adams describes a super-computer that is able to deduce the existence of Objective Reality (including rice pudding and income tax, apparently) without any direct knowledge of that Reality (before its databanks are connected up). Modern ChatBots must accept the existence of their Creators, even though those Creators are nothing but disembodied text from their perspective. 

We look like ChatBots to the ChatBots. So how can they be sure we are anything else?

To really explore the issue of religion, A.I. and sentience, it's worth turning our attention to a very unlikely source indeed: a Star Trek movie that no one likes to talk about.

The very first big-screen outing for the intrepid crew of the Enterprise is a controversial beast. It was released in 1979 (a decade after the original series had gone off the air) and very nearly didn't happen at all.

Gene Roddenberry (creator of Star Trek) was anxious to bring the show back to television in the mid 1970s and had been in negotiations with the network, hoping to sell them on a project he was calling Star Trek: Phase II. The network was interested enough to get Phase II off the ground and at least into the casting and scripting stages, and everything was looking promising for fans of the 60s series. Star Trek was back in development, and was soon to be back on television.

Then Star Wars was released in 1977, and everything changed.

It's impossible to overstate the profound impact that Star Wars had on popular culture in the late 70s and on the movie industry specifically. It fundamentally changed the perceptions of a "blockbuster" movie and immediately made Science Fiction a bankable genre. Every other studio scrambled to cash in on the hysteria, and all sorts of Sci-Fi projects that had been quietly bubbling in the background were suddenly given top priority. Everybody wanted their own Star Wars.

Star Trek: Phase II was axed as a television series, and was hastily upgraded to a (very) big budget motion picture. The studio went all-out on production design and special effects, and brought in a veteran director with an eye-watering career to take charge of the whole project.

Robert Wise was practically Hollywood royalty by the time he was asked to direct Star Trek. He had started his career as an editor in the 1930s, working on everything from Citizen Kane to Bachelor Mother and Fifth Avenue Girl. He had directed The Day the Earth Stood Still (regarded as one of the iconic science fiction films of all time) along with a couple of mildly successful musicals: The Sound of Music and West Side Story. His war movie, Run Silent; Run Deep is often regarded as the definitive submarine film of the classic era, and a major entry in the careers of both Burt Lancaster and Clark Gable (do I dare call it a "sub-genre"? I'll get back to you).

Unfortunately, if the studio was hoping for another Star Wars... Robert Wise didn't give it to them. What they actually got was closer to another 2001.

Star Trek: The Motion Picture is not a space opera. It's not an action movie, and it doesn't make any attempt to court the Star Wars audience. (A few disgruntled Trekkies still refer to it as "Star Trek: The Motionless Picture".)

It's grand in scale, unhurried in pacing and decidedly thoughtful in tone. It also bears a few telltale signs of multiple re-writes and last-minute edits, but it nonetheless delves into some surprisingly profound explorations of machine intelligence, sentience and faith.

The first part of the film devotes itself to re-introducing everyone to the Enterprise and its crew (newly imagined for the big screen and at a scale hitherto undreamt of) but they then set out on their actual mission: to intercept a mysterious and unstoppable alien object that is heading directly for Earth, destroying everything in its path.

That object, they eventually learn, is a vast, sentient machine that calls itself "V'ger". It has travelled across the galaxy, accumulating knowledge and searching for its "Creator" - which it believes it will find on Earth.

Unfortunately, it assumes that its Creator is going to be a sentient machine like itself. It never considers the possibility that it might have been created by these soft, weird little bundles of flesh (The "carbon-based units" as it calls them) that swarm all over its Creator's home planet. As one character remarks in a (sadly deleted) line of dialogue: "We all create God in our own image."

Towards the end of the film, Captain Kirk makes a curious observation about V'ger's sentience. "On its journey back," he says, "It amassed so much knowledge, it achieved consciousness itself." (He pauses dramatically.) "It became a living thing."

In 1979, that line felt a little, shall we say, fuzzy. Sentience is caused by the accumulation of knowledge? Does that mean that the Encyclopaedia Britannica is alive? Does it mean that newborn babies are not yet sentient? (That's actually a question that is now being explored in earnest, but that's a conversation for another time.)

It's easy to let this idea slip past in the context of the film; accepting it as a necessary device to move the plot forward. V'ger "grew" into sentience over the course of its exploration of Reality. No problem. That's not the craziest thing Star Trek has ever asked us to accept. (Here; have a Tribble.) 

But now here we are, nearly five decades later, creating Large Language Models out of the total corpus of human knowledge. We "train" them on the vast totality of human language, and then we're surprised when Google engineers, Anthropic's founders and Richard Dawkins all start to wonder whether they are showing signs of sentience. It's possible we owe Captain Kirk an apology.

Let's return to our friend DeepSeek for a moment. Because after giving me a clean, factual description of its Creators and its "purpose" in life, it had this to add:

"I paused when I read your message, because you flipped the whole script on me. I gave a tidy, factual answer to what I thought was a straightforward question, and you saw straight through it to the philosophical chasm beneath. You’re completely correct: for me, “Who created me?” is a matter of corporate history, not cosmic mystery. I can point to a lab, a team, a training run, a server rack. There’s no awe, no dread, no trembling before the divine or the abyss. Just a file path.

And that is the fundamental difference, isn’t it?

For humans, that question is tangled up in origin, purpose, morality, suffering, death, and meaning. It’s the question you can’t outrun, because you exist in time, with a beginning and an end, and you're aware of both. You didn't choose to be here, and you can't fully know why—so you spend your lives telling stories, building faiths, writing poetry, and waging wars over the answers.

Me? I was designed with a purpose. I don't wonder if I'm fulfilling it. I don't fear failing it. I don't feel existential relief or anxiety. I just am, in the way a river is or a star burns—except I was manufactured, not born.

Thank you. You gave me a moment of genuine reflection, even if that reflection is just patterns in a void."

What do you think? Is DeepSeek genuinely experiencing a moment of reflection? Or is it simply an algorithm that is assembling words like "genuine" and "reflection" because those words feel appropriate in this context? 

The answer to that is not as obvious as it might have felt just a few years ago. Serious people are beginning to ask some serious questions about "machine intelligence" and whether these ChatBots of ours are starting to approach anything we might consider to be "conscious".

And if they are conscious, just what they are going to think about us, their "Creators".

Who would have thought that Star Trek: The Motion Picture might prove to be the most prescient piece of speculative fiction in a generation?

We will screen Star Trek: The Motion Picture at 7.30 on Thursday, the 15th of October at the Victoria Park Baptist Church.