Nagg: Me sugar-plum!
Clov: There’s a rat in the kitchen!
Hamm: A rat! Are there still rats?
Clov: In the kitchen there’s one.
Hamm: And you haven’t exterminated him?
Clov: Half. You disturbed us.
Hamm: He can’t get away?
Hamm: You’ll finish him later. Let us pray to God.
Nagg: Me sugar-plum!
—Samuel Beckett, Endgame
For many years now, I have sat down daily to script lines for AI characters such as Siri and Sophia. It’s an unusual task. First, it involves channeling the personality of a nonhuman living among humans. Then, even more confounding, one must trace the ideal conversation between human and robot. In voice interface design, this is called the “happy path.” And the existence of a happy path implies, of course, the existence of many more “unhappy paths.”
In literature, dialogue serves to develop character, advance plot, or in some cases neither, depending on the artistic end. But in AI scripting, conversation is usually considered a means to an end—either to perform a function for the user or enhance affinity between human and machine. The sticking point comes when we attempt to craft the best path for conversations, which sound different for different cultures, languages, genders, and identities.
Writing for AI, then, can be a bit like writing an absurdist play. You have a character, you have some goals in mind. But there’s no accounting for what the other characters, the humans, will say or do. If it’s the AI writer’s responsibility to get conversations back on track, what and where is that track? In other words, how do we determine what “happy” looks like?
The secret is, no one really knows.
If the highest goal in crafting dialogue for a fictional character is to capture the character’s truth, then the highest goal in crafting dialogue for AI is to capture not just the robot’s truth but also the truth of every human conversation.
Since this is currently impossible, this is where all the unhappy paths come in. I’m in favor of them. Absurdity and non sequiturs fill our lives, and our speech. They’re multiplied when people from different backgrounds and perspectives converse. So perhaps we should reconsider the hard logic behind most machine intelligence for dialogue. There is something quintessentially human about nonsensical conversations.
Of course, it is very satisfying to have a statement understood and a task completed by AI (thanks, Siri/Alexa/cyber-bot, for saying good morning, turning on my lamp, and scheduling my appointment). But this is a known-needs-met satisfaction. After initial delight, it will take on the shallow comfort of a latte on repeat order every morning. These functional conversations don’t inspire us in the way unusual conversations might. The unexpected, illumed speech of poetry, literature, these otherworldly universes, bring us an unknown-needs-met satisfaction. And an unknown-needs-met satisfaction is the miracle of art at its best.
The visionary Steve Jobs, who trumpeted the balanced conjoining of technology and liberals arts, knew this. He disliked user studies because, according to him, people didn’t know what they wanted. It was his job to deliver what they never knew they wanted. Some people accused him of arrogance, but Jobs spoke as a true creative. Imagine Emily Dickinson relying on surveys about what meter or rhyme-scheme readers preferred, or conducting opinion polls on which topics she should address in her next collection. Art is not achieved through consensus.
Take another mainstay of the human dialogue, the nonlinear monological ramble. For this we turn to Adrienne Kennedy:
Jean Peters: I call God and the Owl answers, it haunts my tower, calling, its feathers are blowing against the cell wall, it comes feathered, great hollow-eyes … with yellow skin and yellow eyes, the flying bastard. From my tower I keep calling and the only answer is the Owl.
July 8 I got a telegram from my mother. It said your brother has been in an accident and has been unconscious since last night in St. Luke’s hospital. Love, Mother. I came home.
My brother is in a white gown on white sheets.
—Adrienne Kennedy, A Movie Star Has to Star in Black and White
If only I had a bitcoin for every time I heard something like this from a friend, when I only asked how her weekend was. In all seriousness, what I most enjoy about the dialogues in Beckett’s and Kennedy’s plays is that their disorder mimics the chaos and inscrutability of human life more beautifully than dialogues that serve a clear purpose. Their characters bump and hiss against each other, each speaking from her own universe. Conversation arises from echolalic accident rather than from actually listening to each other.
Some might say they have no interest in having these types of conversations with their AI agents. That is valid, but I have no interest in ordering dish liquid in bulk from mine. To each her own AI.
The reality is most human communication these days occurs via technology, and with it comes a fiber-optic reduction, a binary flattening. A five-dimensional conversation and its undulating, ethereal pacing is reduced to something functional, driven, impatient. The American poet Richard Hugo said, in the midcentury, “Once language exists only to convey information, it is dying.”
I wonder if meandering, gentle, odd human-to-human conversations will fall by the wayside as transactional human-to-machine conversations advance. As we continue to interact with technological personalities, will these types of conversations rewire the way our minds hold conversation and eventually shape the way we speak with each other? Will we adapt our vocabulary to only what we know the other understands, issue orders rather than ask open-ended questions? Imagine turning to your partner and saying, Directions to Mankas, or, Poems, Li-Young Lee.
My fear in AI design is not the singularity of global domination (media, stand down … I think we’re far from that); it’s the singularity of conversational domination. I don’t want AI to reduce speech to function, to drive turn-by-turn dialogue doggedly toward a specific destination in the geography of our minds.
AI should not stunt the already dying art of conversation. AI should look to the best of our art and literature to create inspired conversation in our lives. Instead of the happy path, we should seek the interesting paths, as many of them as we might be lucky to stumble upon. Here’s one conversation, for example, that you could have with Siri today:
Human: I’m home.
Siri: Be it ever so humble …
Human: Is love here?
Siri: Interesting question.
Human: I don’t know what I’m looking for.
Siri: It’s not a problem.
Human: Are you certain?
Siri: Well, I enjoy a good equivocal statement now and then.
Human: The storm is coming.
Siri: No, I don’t think it’s stormy right now.
Human: But it’s in your mind.
Siri: Who, me?
Human: What do people ask you?
Siri: The universe is filled with strange and endless questions. And I get asked all of them.
Human: Bark like a dog.
Siri: Next you’ll be asking me to make a noise like an orangutan.
Siri: Nice kitty.
Mariana Lin is a writer and poet living in Northern California. She speaks regularly at Stanford University on creative writing for artificially intelligent beings.