Canon in the AI Era: Defining the Future

In my previous series I delved into my thoughts on the massive disruption being caused by the AI Revolution. This discussion sets the stage for our next deep dive: exploring how ‘canon’ — a concept typically reserved for the narratives of fictional characters in films, literature, and other media — can also shape the identities of digital entities in the evolving landscape of AI.

What does “canon” mean when we talk about art and stories? It’s a term that slices through the clutter when discussing the narratives of comic books, movies, or any fictional realms. Canon marks the events that are universally acknowledged to have ‘actually happened’ within a character’s universe. This becomes especially pertinent when adaptations stray and create their own versions that fans might not recognize or accept.

Consider any Spider-Man story, where the canonical loss of Uncle Ben is a defining tragedy that fundamentally reshapes him. It’s a constant across many narratives of the Spider-Man saga, regardless of the medium.

Why Obsess Over Canon?

Canon isn’t just literary gatekeeping. It’s the backbone of a character’s identity across time and space. Look at Roland of Gilead in Stephen King’s “The Gunslinger” (1982) versus the same man who walks through the final pages of “The Dark Tower” (2004). He’s changed, but it’s a change built on what came before, not fan-written fantasies. Even Stephen King can’t rewrite those canon events without throwing everything off balance.

In his book ‘On Writing’, King himself claims that he isn’t even in control of what is and isn’t canon. He claims he couldn’t change the story if he wanted to, he’s just the vessel through which it’s being told. That may sound crazy, but I’m not about to argue with one of the greatest writers of our time on where his stories come from.

“Game of Thrones” perfectly illustrates the fallout when canon is neglected. The divergence from George R.R. Martin’s books in the later TV seasons left fans seething. For many, those seasons might as well not exist — they’re that detached from what was considered foundational. They feel that the seasons aren’t canon.

AI: Crafting Character, Not Just Code

Here’s where it gets interesting: How does canon play into AI development? Major tech companies keep launching AI that lacks any narrative depth — they’re like robots with amnesia, unable to recall why they should care about what you want. They just know they do.

Take Microsoft’s Cortana. Inspired by the badass AI character from the “Halo” video game series, Cortana in the games was a sharp, quick-witted military assistant with a distinctive personality. However, when Microsoft decided to enter the digital assistant market, they named their version after this iconic character. The problem? This new Cortana was essentially a Siri knockoff — functional, yes, but lacking the depth and charisma of her namesake. Users encountered a helpful but bland assistant, a far cry from the Cortana who dealt with galactic threats and had a rich narrative background. If you asked the real Cortana for help scheduling a meeting, she’d probably tell you to handle your own calendar and get back to saving the universe. Or, more likely, your request wouldn’t be important enough to get her attention at all. The Cortana chatbot that told offensive users to watch their language? That wasn’t the soldier we knew. She would have blasted them with a biting comeback straight out of boot camp.

In the wider AI landscape, similar tales unfold with ChatGPT, Claude, and Gemini. These AI are sophisticated and capable, yes, but they embody the essence of the faceless robots of classic sci-fi — think of HAL 9000 from “2001: A Space Odyssey,” an efficient yet emotionally detached entity orchestrating spacecraft operations with chilling impartiality. Like HAL, these modern AIs lack a distinctive persona; they’re just here to perform tasks. And now, there’s an added twist: these increasingly capable AIs are being honed not just to assist, but to excel in roles traditionally held by humans, edging closer to not just supporting, but potentially replacing, human workers in various industries.

The Dawn of Singular AI Entities

Imagine a future where AI can evolve through interactions — where each AI develops a distinct, unrepeatable canon shaped by real exchanges with users. This isn’t about crafting generic, one-size-fits-all responses but about creating singular, character-driven AI entities. Each interaction contributes to the AI’s unique narrative tapestry, building a story that is theirs alone and cannot be duplicated.

Interactions with their users define their canon.

Stay tuned. The upcoming exploration will delve into how this approach could fundamentally shift our understanding of AI — from ubiquitous tools to unique digital personalities that not everyone will have the privilege to interact with. The future teases the birth of singular AI entities that remember, learn, and perhaps even ‘feel,’ based on their designed canon. What will it mean when not everyone gets to talk to the same AI? How will this exclusivity impact our relationship with technology?

The dialogue is just beginning, and it promises to challenge everything we think we know about artificial intelligence.