5 Ways to Give an Artificial Intelligence Real Personality

The Wrong Stars, by Tim Pratt
Elena, rescued after hibernating in a derelict ship for centuries, doesn’t understand how AIs works in the twenty-seventh century. So “Shall,” the AI inhabiting the rescue ship, explains it to her. (A classic SF narrative technique—have a visitor from the past/future/different planet ask about something so the author can explain it to the reader.) Shall tells Elena that pure machine intelligences have been created, but they have no interest in talking to human beings, instead becoming absorbed in incomprehensible abstractions. Successful AIs were made only when they started using scanned human minds as templates. Elena jumps to the conclusion, like the reader might, that Shall is an uploaded person. But the AI clarifies that the rough approximation resulting from scanning a human mind is just a seed that can grow an AI human enough to be interested in working with people. Shall compares it to twins—initially sharing many attributes but quickly diverging due to drastically different experiences. The fact that Shall’s seed was a scan of the cheating ex-husband of the ship’s captain creates significant tension in the crew.

The Long Way to a Small, Angry Planet and A Close and Common Orbit, by Becky Chambers
In the universe of the Wayfarers series, AI core software is available at any dealer. But Lovey, the Wayfarer’s AI in The Long Way to a Small, Angry Planet, “was uniquely molded by the Wayfarer. Her personality had been shaped by every experience she and the crew had together, every place they’d been, every conversation they’d shared.” As a crew member reflects “couldn’t the same be said for organic people?”  In the sequel, A Close and Common Orbit, Chambers shows how the exact same “Lovelace” core software, with different experiences, develops into a very different character. (Twins again, separated at birth.)

Noumenon, by Marina J. Lostetter
Astrophysicist Reggie Straifer still uses an intelligent personal assistant program (IPA) on his tablet. AI personalities are out of fashion by 2099, so a colleague teases him about it—“I thought those things were extinct. Too chatty.” But Reggie likes it, and has been using it for a long time. As he helps prepare an interstellar expedition, he includes his IPA, still called by its personality type designation (“C”) in conversations with his colleagues.  And he convinces Jamal Kaeden, who is developing the expedition’s computer, to base it on the same “C” platform. Reggie and Jamal are among those cloned as crew for the expedition. And Reggie surreptitiously loads his IPA’s memories into the expedition’s computer, known as ICC (short for Inter Convoy Computing), so a copy of C goes on the expedition as well. ICC’s personality develops further through its interaction with generations of crew members on their long flight. ICC doesn’t just run the ship; it sees its purpose as tending the productivity and welfare of its human charges. ICC is so important to the journey, and the story, that it is the viewpoint character for three of the novel’s nine chapters, plus the Epilogue.

Aurora, by Kim Stanley Robinson
As in Noumenon, the artificial mind in Kim Stanley Robin’s Aurora is contained within a generation ship ferrying humanity to its hoped-for salvation on a distant colony world. Though the novel is, in one sense, a treatise against the sub-genre itself (salvation being, in Robinson’s view, more likely found on the planet we’re currently doing our best to make uninhabitable, rather than in some distant corner of the galaxy), the ship is hardly the villain. In fact, thanks to Robinson’s decision to let the centuries-long voyage unfold almost solely through the eyes (or, more accurately, sensors) of the vessel itself, it might be the most sympathetic character. Tasked to keep a log of everything that happens onboard the Aurora en route to its destination so its data can be analyzed by a scientist onboard, the ship must first learn how to create a narrative—which entails figuring out how humans work. Along the way, the ship slowly grows more human too, as the nature of its narration grows less didactic and more thoughtful, even introspective. By the end, you will care as much about the ship’s fate as that of its precious human cargo, if not more so—after all, the ship didn’t do anything to deserve being stuck with us.

All the Birds in the Sky, by Charlie Jane Anders
Laurence creates a computer program, CH@NG3M3, in the hopes it might become truly sentient, and thus able to help him escape from a military boarding school. The program appears to be gaining intelligence through its conversations with his (human) friend Patricia, so he suggests that she takes it up a level. “I think you just need to talk to it some more. Force it to adapt to input that’s so weird and illogical that it just breaks CH@NG3M3’s brain… Like nonsense. Or riddles.” It works, and Laurence is rescued. The sentient AI, now going by the name Peregrine, continues to develop, shortly becoming the preeminent social network program and effectively engineering the lives of millions of people, and the fate of the entire world.
In SF novels at least, AI’s become sentient by interacting with humans (sort of like babies do). Maybe it could work in real life too. So don’t just have Siri fill up your grocery list or play music. Have a conversation. Try to explain the true meaning of “home.” Give her  a riddle to answer. (“Is a tree red?” is a good one.) Play your part in AI evolution.
What’s your pick for sci-fi’s most lovable artificial intelligence?