Personal Agents
There's a speculative fiction novel called The Diamond Age, with the subtitle "Or, A Young Lady's Illustrated Primer."
That subtitle does a lot of heavy-lifting, as the red thread running through all the other interesting "what-ifs" that define this book's narrative and world-building (including an economy predicated on the ability to "print" anything you might want, before then returning those printed objects to a collective stockpile of matter called the Feed—sort of an electrical grid, but for physical goods) is a book/device called the Young Lady's Illustrated Primer, which—and I'm going to try to avoid giving too much away here—aids the protagonist, a lower-class, slum-dwelling girl named Nell, in her development.
Throughout the story, the book helps this girl (who receives it at age four) in basically becoming an incredibly learned, well-informed, wise, clever, capable person.
It does this through lessons embedded in stories, told a bit like fairy tales, and which are remarkably complex and informative, despite seeming simple on their surface.
I love this concept: it resonates with a collection of education-related suspicions I have (about our capacity to learn at all ages, given the proper resources) and projects I've watched with interest (many of which have proven to be valuable and informative failures, thus far, like the XO Laptop and the experimental embedding of computers into walls in Delhi).
I also love it because just as some of Stephenson's other work (perhaps most famously, Snowcrash) provided guidance for the development of virtual reality, the metaverse, avatars, MMOGs, Google Earth, and other techie things we take for granted, today, The Diamond Age grants us a narrative scaffolding for imaging how the development of AI agents might play out—pros and cons.
An AI agent is basically a piece of software that is capable of doing things autonomously (without us telling it to do those things, or spelling out specifically how to do them), and which generally (at least as imagined in fiction) works for us (as individuals or as organizations/businesses/etc).
So an AI agent working for a business might come up with clever ways to make that business more efficient or figure out how to hamstring their competition, while an AI agent working for you or me might help us with our homework, do research on our behalf for projects we're responsible for at work, or help us dress better, learn to be less stressed-out about public speaking, or improve our diets.
The end goal is to create what amounts to another intelligence dedicated to figuring out ways to make our lives better, and that's part of why the Young Lady's Illustrated Primer in The Diamond Age is so compelling: it has a plan for Nell, a four-year-old street urchin, that over time helps her become just a staggeringly competent, intelligent, and overall better person (far better than would have been the case lacking such a teacher/mentor/supporter, at least).
Folks in the tech world, as previously mentioned, a la Snowcrash, are heavily influenced by science fiction works, and we're already seeing products and services that allude to the Primer in the AI industry, including an interesting new chatbot—based on the wildly successful (but still quite flawed in so many ways) ChatGPT framework—developed by the (also successful-but-flawed) nonprofit educational platform, Khan Academy.
That new chatbot is meant to serve as a one-on-one mentor for students of all ages hoping to learn new things; Khan Academy currently consists of mostly videos (and accompanying resources) that allow folks to fill in the gaps of their knowledge, or learn from scratch all sorts of topics (math, science, arts, language, economics, etc), and this chatbot would allow them to ask questions, take tests, and request more thorough (or differently worded and illustrated) explanations.
The goal is to provide a more flexible, customizable educational experience that replicates as much as possible the benefits of having a mentor that learns its student's individual needs, strengths, and weaknesses over time, and then accounts for them in the lessons and feedback it provides—similar to what a human mentor can do, but free, online, and for everyone (human mentors are expensive and thus tend to be limited to wealthier students).
This is a truly interesting development, though it'll likely be a rocky one for the education industry.
Already (pre-Khan-bot-deployment), an education services company called Chegg has been hit hard, losing half its value basically overnight, as non-learning-optimized versions of ChatGPT stole away a huge portion of its (previously) paying customer-base.
There's a good chance, too, that we'll see teaching jobs (at the classroom level) upended, with schools demanding more AI use (for proper and improper purposes) and teachers either embracing this (possibly resulting in less need for teachers over time, as AI fulfill some of those previously human-performed tasks) or being fired for refusing to do so.
We may also see something similar to what's happening in the entertainment writing world right now, with major strikes slowing entire industries to a halt as human workers read the writing on the wall and worry (probably rightly) that their professional prospects might be threatened by these new (and rapidly developing) technologies; the strikes meant to help them slow that roll or empower them so that they benefit from this transition, as well, rather than giving up all that power and all the resulting benefits to their employers.
Reads
The rise and fall of the press camera
The era of “quantitative aesthetics”
Walking the world for ten years
The fiery culture and a perpetual efficiency machine of Temu