My copy of Kazuo Ishiguro’s Klara and the Sun has a blazing red dust jacket. Current events and the eye-catching spine recently prompted me to pick it up again.
The novel takes place in a near-future where artificial intelligence and robotics have advanced considerably. Klara, a solar-powered “artificial friend” or “AF,” tells her own story. The glimpses of the wider setting we get from her narrative are mostly recognizable, aside from the occasional hint of disquieting technological and social developments. As in Ishiguro’s Never Let You Go, much of the world-building occurs in the reader’s imagination. In the case of Klara and the Sun, the focus is squarely on Klara’s recounting of her experiences and her interactions with the humans around her.
In Klara’s world, a risky gene-editing procedure called “lifting” is common for more advantaged children. Society is highly segmented, with genetically boosted children being educated remotely by tutors and attending elite colleges only for those who have been “lifted.” AFs provide support and socialization for these increasingly isolated children as they march towards adulthood. The wrenching decision of whether or not to “lift” a child reminded me of current parents’ struggles to manage their children’s use of technology.
For such a short and readable novel, Klara and the Sun raises profound questions about childhood, the role of technology in human life, and the ethics of creating artificial intelligence. As interesting as these questions are, Klara steals the show. Her narration is patient, transparent, and reliable. She was designed to be observant and emotionally intelligent– the better to serve her young human companion Josie, who is suffering the possibly terminal side effects of the “lifting” process.
When I realized that Klara was an artificial being, I tried to position her in the panoply of androids and intelligent computers I’ve met in science fiction– and more recently through my own use of AI such as ChatGPT. Klara never describes her own physical appearance, but we learn that she could pass for human (no dials or glinting metal like C-3PO). She is intuitive, intelligent, and observant with human-like feelings and even suspicions. Like Star Trek’s Lt. Commander Data, Klara has a rigid sense of mission, protecting and nurturing her human companion. Unlike Ex Machina’s Ava, she is utterly selfless and lacking in duplicity. Klara is eminently practical and responsible when it comes to her sickly charge. But her irrational reverence for the Sun, the source of her energy, is almost religious and sets her apart from other fictional artificial beings I've encountered. The complete faith she puts in the Sun feels naïve. Was she designed this way or is her religiosity a byproduct of her sentience?
When I first read Klara and the Sun, large language models (LLMs) such as ChatGPT had not yet been released. Four years later, AI is emerging as the most important technological development of our time. Although LLMs work by digesting vast quantities of human language, morality and faith like Klara’s don’t appear to be an inherent property of these systems. Boosters of “human-centered AI” advocate for AI developers to build human values into their systems and create AI that can support human endeavors rather than replacing them. Klara’s physicality obviously sets her apart from ChatGPT, but her ability to flawlessly mimic her human counterparts is not so dissimilar from today’s nascent AI offerings. I was surprised by how much trust Klara inspires in the humans around her, even when her ideas are irrational. This idolatry of technology feels familiar– we tend to venerate our tech leaders and too easily trust the results of a Google search (or more recently, an AI query).
As it becomes clear why Klara was really brought into Josie’s life, deeper questions arise. Can Klara replace a human being? Is it ethical for that to be her purpose, even if she proceeds willingly? Initially, she believes she could successfully become Josie– the ultimate Turing Test. Klara’s explanation of what it would take to “know” a human deeply enough to become them is fascinating:
‘The heart you speak of,’ I said. ‘It might indeed be the hardest part of Josie to learn. It might be like a house with many rooms. Even so, a devoted AF given time, could walk through each of those rooms, studying them carefully in turn, until they become like her own home.’
Josie’s “heart” is a finite problem that Klara believes she is capable of solving if that is what is asked of her. But she doesn’t seem capable of thoroughly considering if this is actually the best choice for those around her. Klara’s only agency is in the service of others (here her actions actually demonstrate great creativity). I wondered if her selfless nature was part of the AF design– a sort of failsafe against a “Skynet” moment– or if it could actually be “human-centered AI” taken to the extreme. But is it ethical to create a being that is so selfless? Is it possible to be sentient and have no self-agency? Or is Klara actually the ideal human?
Klara and the Sun is an achievement. It is a compelling story told in clear language with just the right amount of science fiction detail to stimulate the imagination. Ishiguro manages to create a cohesive world that feels at turns familiar and ominous. And his cool-headed and warm-blooded narrator raises complex questions that might be relevant sooner than we think.
Comentarios