Nobel Laureate Kazuo Ishiguro probably didn’t intend his quietly intense, metaphysical science fiction novel, Klara and the Sun, to be a didactic tale for Data Science workers. For those of us professionally orbiting or embedded in Artificial Intelligence and Machine Learning, Ishiguro’s haunting story subtly addresses many of the challenges and controversies in the field today.
Klara, an “Artificial Friend” (AF), is the eponymous narrator of this tale. AFs exist to provide unfailingly considerate and loyal companionship for the well-heeled families who can afford them. The novel begins with Klara and her brethren in the downtown AF store awaiting purchase by their would-be charges.
Eventually, Klara is adopted by Josie, a 14-year-old with an unspecified illness. Klara becomes Josie’s friend but not equal.
Ishiguro’s novel takes place at an indeterminate time in the near future, somewhere in the United States. Society is stratified. In-person public education is no longer provided for children in society’s upper echelons. Children learn from their “oblongs,” digital devices, with a direct connection with their tutors. Planned get-togethers with other “homeschooled” children to promote socialization have a Lord of the Flies vibe as parents hover on the periphery, letting their maladjusted kids resolve their differences without adult intervention. Ishiguro finished writing Klara and the Sun a few months before COVID-19 hit, making him not just profound but prescient.
Here are a few of the controversial Artificial Intelligence topics addressed in Klara and the Sun:
- Unsupervised learning. Klara is an unreliable narrator because her inferences arise from her narrow worldview. Ishiguro presents Klara’s perfectly logical but off-kilter insights with hilarious brilliance:
The Friend’s Apartment was inside a townhouse. From the window of its Main Lounge I could see similar townhouses on the opposite side of the street. There were six of them in a row, and the front of each had been painted a slightly different color, to prevent a resident climbing the wrong steps and entering a neighbor’s house by mistake.
This is Klara’s understanding of graffiti:
…saw two gray buildings side by side that weren’t the same height, and someone had made a cartoon painting on the wall of the taller building where it stood above its neighbor, perhaps to make their discrepancy less awkward.
Klara’s misconceptions take a dark and potentially dangerous turn, in part because she’s reluctant to get a reality check from humans.
In his New York Times article, author and magazine journalist Nick Bilton worries that AI’s pragmatically logical solutions to human problems may result in the elimination of humanity:
But the upheavals can escalate quickly and become scarier and even cataclysmic. Imagine how a medical robot, originally programmed to rid cancer, could conclude that the best way to obliterate cancer is to exterminate humans who are genetically prone to the disease.
Ishiguro’s presentation is gentler than Bilton’s cataclysmic predictions. Still, Ishiguro’s story portends how an AI might use its brutal machine logic against humans.
2. Explainable Artificial Intelligence/Machine Learning. AI Practitioners face pushback because of black-box AI applications. Critics argue, quite reasonably, it’s hard to trust an AI’s conclusions when there’s no lens into its decision-making. Similarly, Klara experiences hostility from humans suspicious of AFs’ motives. The aggression is explained to Klara like this:
Klara, the fact is, there’s growing and widespread concern about AFs right now. People saying how you’ve become too clever. They’re afraid because they can’t follow what’s going on inside any more. They can see what you do. They accept that your decisions, your recommendations, are sound and dependable, almost always correct. But they don’t like not knowing how you arrive at them. That’s where it comes from, this backlash, this prejudice.
Explainable AI is an emerging field that attempts to provide “how the sausages are made” visibility into AI/ML pronouncements. If a human understands the AI/ML decision-making process, it’s easier to make tweaks, providing better results akin to a math teacher demanding to see “the work” along with a final answer.
3. AI takeover of human jobs. AlthoughKlara works alongside humans, she experiences hostility from humans about AI-related job takeovers. When Klara and Josie attend a play, a fellow theatergoer makes a dig against AFs that may also reflect Ishiguro’s subtle disapproval of today’s anti-immigration tropes:
First they take the jobs. Then they take the seats at the theater?
Taiwanese-American computer scientist, businessman, and writer Kai-Fu Lee believes job loss due to Artificial Intelligence is inevitable. He argues the safest jobs involve high degrees of emotional intelligence — the hardest to replicate in Artificial Intelligence. Lee explains his thinking in this blog post:
AI has infiltrated the Klara and the Sun society so deeply that knowledge workers have been “substituted” by AIs. The role of AI in this society has likely increased societal stratification instead of reducing it. One of the characters in Klara and the Sun, a highly skilled engineer who now sits toward the bottom of the hierarchy, puts a brave face on being replaced:
Honestly? I think the substitutions were the best thing that happened to me. … I really believe they helped me to distinguish what’s important from what isn’t. And where I live now, there are many fine people who feel exactly the same way.
Arts and literature often provide glimpses of the future. Although a great deal of Science Fiction, including Klara and the Sun, presents an anti-utopian society, a future of suffering and injustice is not a foregone conclusion. Technologists may avoid a dystopian future by absorbing these lessons from literature and heeding these powerful technologies’ ethical and practical considerations.