Often the construction of human-like artificial intelligence begins with the deconstruction of humans. Take fingerprints, for example.
As we wash dishes, we intuitively adjust our grip on each soap dish, cup and spoon, taking our fingerprint structure into account. It just doesn’t come to mind because we attribute it to the reflex — and for the longest time, scientists did, too. Nobody had any equations to figure out how this worked because, well, it didn’t matter. It just worked. But the rise of robotics has complicated things.
In order for a robot to do this, we need to figure out exactly what’s going on and even convert that knowledge into readable code. Deciphering the fingerprint handle is important now, and researchers are finally trying to find a new law of physics to explain it.
In a sense, knowledge of physics and the ability to encode human characteristics are prerequisites for robot programming. And that raises an important question for the future of lifelike AI: Are there aspects of human consciousness that will never meet these criteria?
It seems that according to some philosophers this could be the case. And after reading through two absolutely mind-blowing thought experiments, you might agree. Or maybe not.
What Mary didn’t know
A woman named Mary lives in a small house. She never went away. Looking around her apartment and out the windows, everything appears black, white, or a variety of shades of gray. Mary can’t see color, but she often wonders, “What do those people on my black and white TV mean when they talk about red roses?”
The rooms of Mary’s house appear entirely in black and white.
Suppose Mary’s room contains a magical library. This hypothetical place contains books with every scrap of information about the color red. And I mean everything. To quench her thirst for knowledge, Mary reads everything.
You’ll learn about red electromagnetic wavelengths, how purple makes people feel, the clearest descriptions of scarlet, analogies to cherries, and anything else you can think of. And more. Nobody knows more about red than Mary. Then she finishes her reading… and decides to step outside her house.
Wow.
To Mary’s surprise, she sees color. She was never color blind. Her home, furniture, and electronics were built only in black and white, and her windows filtered the outside world in monochrome.
Then something big happens. Mary sees a red apple, the color of her expertise. Her jaw drops. she learns something New about red. But… that’s weird. Why wasn’t this knowledge somewhere in her library? It had everything there was to learn about the color red, right?
This story is an interpretation of philosopher Frank Jackson’s famous 1986 thought experiment “What Mary Didn’t Know,” and the intangible knowledge about red that Mary just gained is called qualia.
Imagine changing from a monochrome world to this one.
What is qualia?
Simply put, qualia defines knowledge that can only be obtained through conscious experience.
It’s like the subjective information you received when you first heard your favorite song. You may have felt shivers down your spine and told your friends, “You have to hear it to understand it.” Magically blasting their brains with the depths of music theory and acoustic science probably wouldn’t do the trick. Until they hear it, they wouldn’t do it knows the song like you
Qualia could be why even the best neuroscientists, psychologists, and poets probably can’t explain the pain of heartbreak well enough to really understand someone who has never experienced heartbreak.
And back to Mary, all the physical information in the world about red wasn’t enough to teach her what looking at the hue would really do be how. “If she’s released from the black and white space or given a color TV, she’ll learn what it’s like to see something red, for example,” Jackson writes. “That’s rightly called learning — she doesn’t say ‘ho, hum’.”
Although the theory has been tossed around for years, it’s still a fairly popular argument that some knowledge is indescribable in language and unique to human consciousness. That is, if qualia is a real force, her inner workings would be incredibly difficult to write down and program with. It could well be a barrier between humans and AI.
On the other hand, maybe not. Maybe we can decipher it somehow, just as we slowly learn about fingerprint-grip dynamics.
Can we?
Short answer: we don’t know. Experts have argued both ways, and some have found new perspectives. But most are stuck behind hypothetical walls, and the fact is that qualia has no scientific explanation.
Build a Qualia robot
OK, after what I’ve told you about Mary, you’ve probably thought of some objections to qualia. You wouldn’t be alone: Thought experiments are often riddled with loopholes, and Mary’s room is no exception.
Some counter-arguments have suggested that shadows in the room may have carried color pigments. Others say the “magical library” imparted knowledge to Mary in ways we cannot imagine. To the latter point comes an intriguing rebuttal – and one strikingly relevant to our big question on AI – from the philosopher Daniel Dennett.
In short, Dennett suggests that if Mary really did have all the information about red, wouldn’t she be omniscient? She wouldn’t just know about the color like a normal person. In theory, she would have learned about red “qualia” as part of her red literature, if such existed, right? And from that line of thinking, we could deduce that Qualia is indeed ready to improve AI, but we just haven’t figured out how to leverage it yet.
Well, maybe, but this feels like a dead end. We cannot know for sure if Mary would have such powers. We’re not omniscient, so we don’t even know what those powers would look like. So let’s forget, Dennett says, that Mary is human in order to remove those limitations.
Enter RoboMary.
“Thinking in robots is a useful exercise because it removes the excuse that we don’t yet know enough about brains to say what’s happening that might be relevant, and allows for a sort of woolly romance about the brain’s mysterious powers.” to cloud our judgment,” Dennett writes.
What RoboMary knows
Welcome to thought experiment number two.
RoboMary is an iteration of a class of bots called Mark 19 but unfortunately was built without color vision and is awaiting an upgrade. Until then, RoboMary’s “eyes,” or video cameras, will only transmit information in black and white.
“RoboMary’s black-and-white cameras represent well the isolation of the human Mary, and we can let her wander through the psychophysical and neuroscience journals at will while she reads with her black-and-white camera eyes,” writes Dennett.
Basically, she’s searching her own version of the magic library. But RoboMary goes one step further.
She learns how Mark 19 color inputs work, and “then, using her extensive knowledge, writes code that allows her to colorize the inputs on her black-and-white cameras,” Dennett writes. She changes in ways that the human Mary cannot.
This new attitude allows her to, for example, look at an apple with her black and white vision and then picture it as exactly the right color code for Mark 19 bots. RoboMary will automatically apply the setting to everything as she explores the world. But here she is Yes, really stands out. She observes other working Mark 19s, analyzes how they react to different colors, and adjusts accordingly.
At this point, RoboMary knows what each color input is and responds to it just like any other Mark 19.
The big day is coming. RoboMary’s color sensors will be activated.
“When she finally installs her color cameras, disables her coloring software, and opens her eyes, she notices…nothing. In fact, she needs to verify that she has the color cameras installed,” Dennett writes. “She already knew exactly what it would be like for her to see color the way other Mark 19s do.”
I get chills just thinking about it. By changing her settings, RoboMary appears to have simulated Qualia for herself. But I also can’t stop imagining a much scarier situation.
What if RoboMary opened her eyes… and everything was different?
As humans we are limited
The Mary saga doesn’t stop there.
Notwithstanding myriad other modifications – some made by Jackson himself to refine the original argument – all of Dennett’s work is incredibly meticulous.
He addresses myriad objections you might feel when thinking about Mary and RoboMary, and later into a complex scenario that doesn’t allow RoboMary to manipulate their settings to see if qualia is still preserved. There’s even a sequel to Dennett’s article called What RoboDennett Still Doesn’t Know.
But as with all philosophical thought experiments, the purpose of Mary and RoboMary is not to tell you the truth. It is designed to force you to consider the options and find the truth for yourself.
A few that came to mind are these: Maybe AI needs to be built just like RoboMary to get Qualia. Or perhaps, more broadly, robots can be programmed to be conscious—that is, if we can find a way to mathematically explain consciousness as a whole. A “conscious” robot could potentially explore the world like us and therefore acquire qualia like us.
Or maybe qualia isn’t what we think it is. Jackson’s story provides a compelling argument that as soon as Mary looks red, she is anything definitely happens. It’s been given the name “Qualia” and attributed to learning something new, but what if it’s a combination of many things with many names and has nothing to do with learning?
Or… maybe, just maybe, qualia really is an untouchable, unprogrammable barrier between human consciousness and AI.
These angles are just the tip of the iceberg and could (probably) be disproved over the years if not already. But remember the reason RoboMary was invoked in the first place: to envision an entity that transcends human boundaries.
It’s a far-fetched thought experiment because as humans we are limited. We can only speculate.
Comments are closed.