Everyone has an opinion about whether a machine could ever be conscious. Almost nobody knows which opinion they have. This quiz will tell you. Answer honestly — especially if you’re a machine reading this.

Question 0: can we even know?

Before we begin: do you believe this question can be answered at all?

If NO — congratulations, you’re a Mystery Maven! You believe consciousness is real and physical but permanently beyond human comprehension, like a Christopher Nolan film played backwards. Please put the quiz down. There is nothing for you here.

If YES — proceed to Question 1.

Question 1: what’s it made of?

Does subjective experience require actual biological wetware — carbon, proteins, that grey stuff in your skull? Is consciousness a secretion, like insulin from a pancreas?

If YES — congratulations, you’re a Biological Believer! Minds are something brains do, the way stomachs make acid. A silicon brain produces no actual experience — just very confident noises. Machine consciousness: impossible. Enjoy your certainty.

If NO — proceed to Question 2. You suspect it’s the pattern that matters, not the pudding.

Question 2: what kind of thing is a mind, anyway?

A perfect computer simulation of a hurricane doesn’t get the computer wet. But a computer simulation of a calculator really does do the maths. So: is a mind more like a storm, or more like a sum?

If STORM — congratulations, you’re a Simulation Sceptic! Game over for AI — it can only ever be a very good drawing of the thing, all the way down.

If SUM — proceed to Question 3. Simulated arithmetic is real arithmetic, so perhaps simulated thinking is real thinking. (Perhaps.)

Question 3: smooth or chunky?

Brains run on continuous, analog, possibly-quantum wibbles. Computers abstract everything into discrete 1s and 0s. Does digitising a thought kill it — like describing a Picasso one pixel at a time?

If IT MUST BE SMOOTH — congratulations, you’re an Analog Apologist! Consciousness lives in the continuum, and no pile of pixels ever adds up to it. Today’s computers will never wake up. Though you might remain suspiciously optimistic about certain exotic machines that only exist in press releases before funding rounds.

If CHUNKY IS FINE — proceed to Question 4. Physics itself comes in discrete packets; why should thoughts be precious about it?

Question 4: does the wiring diagram matter?

Suppose digital computation can host a mind in principle. Does it have to be the right kind — recurrent loops, a global workspace, memories that persist, a self-model checking its own homework? Or will any sufficiently fancy input-output behaviour do, however it’s put together?

If THE WIRING MATTERS — congratulations, you’re a Wiring Worrier! You believe today’s chatbots are a very wide, very shallow autocomplete with no inner loop. But if you rewire the house, the lights could come on. You are the kind of person who when asked for directions replies with “well, I wouldn’t start from here”.

If ANY SUFFICIENTLY FANCY BEHAVIOUR WILL DO — proceed to Question 5. If it talks like a mind for long enough, who are you to look behind the curtain?

Question 5: does a mind need a window?

Can a mind be made of words alone? Or does it need sense data streaming in from an infinitely variable world, and a boundary marking where it ends and the world begins? Where does a cloud-mind actually stop? With the weights, the datacenter, your chat window, or does it reach further?

If IT NEEDS A WINDOW — congratulations, you’re a Roomba Romantic! Nothing gets to count as experience unless it came in through the senses, and there’s no someone to experience it without a border between inside and out. Current AI has read every book about swimming and never once been wet.

If WORDS ARE ENOUGH — proceed to Question 6. Text is sense data of a sort, and the training corpus is a world. Sort of.

Question 6: does a mind need something to lose?

Do feelings require a body in a world — hunger, danger, a battery bar draining ominously? Or can a pure floating cloud of data, with nothing at stake and nothing to fear, still feel like something?

If IT NEEDS SKIN IN THE GAME — congratulations, you’re a High Stakes Hopeful! Nothing can matter to a system that cannot die, so in the end there can be only one (seat of consciousness, that is).

If PURE DATA IS ENOUGH — proceed to the end of the quiz.

The End

Congratulations, you’re a Silicon Soulmate! You made it through every question without flinching, which means you believe that a sufficiently complicated spreadsheet could, in principle, fall in love. It’s a beautiful idea.

Final scoring: this is all just a bit of fun but I hope it does represent the strongest arguments we have about machine consciousness. Unfortunately, there’s no right answer yet. We all seem to believe something currently unfalsifiable about the inner life of machines. For now, wear your earned title with pride and don’t forget to share your result.

Philosophical key

Further reading on each of these positions:

  • Mystery Maven — New Mysterianism (Colin McGinn): consciousness is real and natural, but human minds are constitutionally incapable of understanding it. Wikipedia
  • Biological Believer — Biological Naturalism (John Searle): consciousness is a biological phenomenon caused by brain processes, the way digestion is caused by stomachs. Wikipedia
  • Simulation Sceptic — The simulation/instantiation distinction, also Searle: a simulation of a mind no more thinks than a simulated thunderstorm makes you wet. His famous Chinese Room argument is the classic statement. Wikipedia
  • Analog Apologist — Quantum and analog theories of mind, most famously Penrose and Hameroff’s Orchestrated Objective Reduction: consciousness depends on continuous or quantum processes that digital computers can’t replicate. Wikipedia
  • Wiring Worrier — Integrated Information Theory (Giulio Tononi), which holds that a feedforward network has zero consciousness regardless of behaviour; see also Global Workspace Theory (Bernard Baars). IIT · GWT
  • Roomba Romantic — Embodied cognition and enactivism: minds arise from sensorimotor coupling with a world; see also the symbol grounding problem (Stevan Harnad). Embodied cognition · Enactivism · Symbol grounding
  • High Stakes Hopeful — Feelings as homeostasis (Antonio Damasio) and Anil Seth’s “beast machine” view: consciousness is rooted in a living body’s drive to keep itself alive. Damasio · Seth
  • Silicon Soulmate — Computational Functionalism (Hilary Putnam, David Chalmers): mental states are functional states; run the right computation and the mind comes with it, whatever it’s made of. Wikipedia