Nowadays, I think about how much the idea of such novelty has shifted in the world of artificial intelligence. Some of the earliest examples of this in education was using genAI to create short skits with the names of your students as the speaking characters, or a poem with your personal favorites referenced in verse. All of this can be created in seconds with a prompt of a few sentences. But the AI rocket ship left that simple spacedock of short skits and poems a long time ago. The potential output media is becoming more dense and complex. Like your name in an original song, complete with music and lyrics? Easy with AI (although you may have to deal with some name mispronunciation). Want to include yourself in a picture with famous people in any setting? No problem. Would you like to make a video inserting yourself in the middle of your favorite movie? Increasingly simple, and more realistic every passing day. But the novelty of such outputs wears out quickly. In the world of AI, novelty is a rapidly diminishing commodity.
One thing that I wonder about is the unintended consequences of this on reading literacy. You've been able to make personalized printed books with your children as the main characters for years if you had the time and money, but now even graphic novels could be composed around "you" within seconds. Of course, fictionalizing your own personal narrative is fine, and there is nothing inherently wrong with using personalized novelty-powered AI to motivate a student. Additionally, even though I listened to my Captain Zoom album more than I want to admit, my musical taste still grew and matured into non-personalized fare. However, with the ease of such machine-enabled personalization, I fret if students will fixate on such parlor tricks at the expense of non-AI options. Will the most popular media we consume tomorrow only be ones with us as the starring role?
Rudine Sims Bishop first talked about books being mirrors, windows, and sliding doors back in 1990:
These three reading opportunities each have their place, but the danger can be in children who, as Bishop says in the video, "always find [only] their mirrors in the books, and therefore, get an exaggerated sense of their own self-worth and a false sense of what their world is like." (To be fair, Bishop originally used this metaphor to explain what can happen when a dominant group of students only see themselves in the majority of their texts, instead of benefitting from a diversity of perspectives.) Mirrors can affirm, but they can also make us Narcissus starting at our own reflection in the pool. The early tech novelty of a mass consumerism product like Captain Zoom has evolved into instant, personalized, present-day AI naval-gazing.
Becky Keene, in her book AI Optimism: A Guide for Redefining Artificial Intelligence in Education, acknowledged "[t]he potential for [AI to create] information bubbles. This is a phenomenon where hyper-personalization leads to learners receiving information from a narrow set of sources that align with their existing beliefs and preferences. This can prevent them from being exposed to divergent values, perspectives, and experiences that are fundamental for their intellectual and personal growth. . . . Instead of solely focusing on hyper-personalization, AI systems should be designed to encourage listeners to explore a broader range of topics and perspectives" (118). I couldn't agree with Keene more, but the operative word here is should. I am not sure how often history's megacorporations have demonstrated themselves to be cautious, sober and liberal-minded leaders who choose people over profit (unless and until faced with significant external pressure, legislation, or bad press), nor how much they are incentivized to broaden our differing perspectives instead of weaponizing them. Perhaps through our thoughtful interactions with the machine, we can shape it to be so; a more pragmatic goal may be to teach our students and ourselves to be more discerning when, whether or if we use AI in the first place, much less when judging its output.
Are there AI tools that pragmatically serve teachers and students without seeming like a sophisticated version of Captain Zoom? I talked in a previous post about the potential for AI to infringe upon creativity, and as we design quality instruction, we should ask deeper, more vibrant human tasks of our students than the low-rigor ones that are often the hallmarks of traditional teaching. That aside, there are some AI tools that uses artificial intelligence for a greater (and useful) good, particularly when it comes to reading literacy. To take but two examples:
- MS Reading Coach requires Microsoft accounts, but it is free, and students can access teacher assignments through a URL or join code. The platform allows teaches to take either already created readings or customize their own, then assign them to students so they can read them aloud; the AI assesses their fluency, and provides a data dashboard for both the student and the teacher on progress and growth.
- While LitLab does allow some personalization of the "decodable books" it can create with the help of AI, I am more appreciative of its various exportable options such as Google Docs or even printing out hard copies, so that students can benefit from teacher-designed output even if they don't have a screen. (While LitLab has a free tier, it also has several premium features, such as oral reading fluency practice.)
In closing, I need to stress that meeting students where they are at, or incorporating students' interests when scaffolding, differentiating, and personalizing instructional materials, are the hallmarks of best practice and universal design. I am simply pointing out the concerns of educators like Bishop and Keene as we determine our next steps with AI. Before we zoom into the hyperspace of hyper-personalization, let's make sure we properly plot our coordinates.
