The edtech market is filled with a variety of tools designed to improve children’s literacy skills, from e-readers to apps to digital libraries. In recent years, a growing number of literacy tools have used generative AI to accelerate children’s reading comprehension and further stimulate their interest in reading.
Recently, a new type of tool has appeared. These tools, called AI-powered reading coaches, assistants, or tutors, use generative AI to provide learners with personalized reading practice, stories, feedback, and support.
Some of these tools focus on specific learning objectives or thematic areas within the story, such as phonics instruction. Some incorporate personal data, such as the child’s name, and offer options to select settings and avatars, giving each child a unique story.
As a professor of reading and child development who specializes in children’s digital tools, I’ve been researching what works and what doesn’t when teaching children to read. Additionally, by collaborating with colleagues through WiKIT, an international research organization focused on evidence in educational technology, we have reviewed multiple tools using generative AI to teach children to read and write. For example, I’ve seen many companies have the potential to create learning breakthroughs by providing personalized fluency exercises and feedback tailored to each user. However, there are very real concerns about the impact these tools have on children’s literature and literacy experiences.
Potential opportunities and drawbacks
Depending on the tool, these AI-powered reading coaches, assistants, and tutors include different elements to support children’s literacy skills. Common features include using voice recognition technology to listen to children read books, using AI to choose from a series of interventions and feedback, and using AI to create narrative texts for children to read. This includes generating prompts and creating personalized prompts based on the child’s abilities. And like many edtech tools, it’s common to use reward systems, such as allowing learners to collect badges and prizes based on their learning progress. Each of these elements comes with its own opportunities and drawbacks.
Using voice recognition technology to listen to your child read or using AI to provide feedback can be helpful as long as the technology is based on a science-backed design. The problem is that many tools claim to be science-based, but have not actually been developed by learning scientists or tested in rigorous assessment studies. Such tools are typically designed to interest and motivate children to interact with the story, but they do not necessarily lead to improved reading comprehension in children.
The same goes for AI-generated stories. They typically engage children by allowing them to make choices, such as what characters and settings to choose for the story, and by personalizing the experience, such as making the main character a character with a personality. child’s name and age; But AI-generated stories often don’t match what science recommends for children’s literary experiences. For example, AI-generated narratives often have inconsistent story elements. On one page, the main character shows her as a five-year-old blonde girl, but on the next page, with no prior indication of time in the text, she transforms into a teenager. Discrepancies in story events are also very common. In a story I recently created with her one of these tools, the main character Natalia (named after her own name, of course) suddenly finds herself in a situation where her new character “Remi’s dog” has nothing to do with her. Now we can have a dialogue. An advance mention of how Remi and the dog came into the story. According to research This kind of narrative confusion confuses young readers and prevents them from empathizing with the characters.
Being research-based is valuable for effective content and narrative text formats. Currently, many stories generated by AI resemble illustrated e-books rather than digital picture books. Illustrated e-books typically only have text drawn to reflect the information within the text. If the text says “Natalia is wearing a yellow shirt, standing in the garden and smiling,” the character will be drawn exactly to match that description. in contrast, High quality children’s picture books, both pictures and text contribute to the depth of the story, broaden children’s horizons, make them think abstractly, and make them ponder. It is a kind of literary experience where poetry paints a picture in the reader’s mind and elevates the reading experience to an art form, which writers like Jacqueline Woodson have achieved in her book Brown Girl Dreaming.
And in high-quality digital children’s books, narration goes beyond simply reciting the written text and expands on the story with additional emotion and drama. The complementary and mutually enriching roles of images, text, and narration in the story children can You can not only become a better reader, but also develop stronger writing skills and media abilities.
Although the aesthetic quality of AI-generated stories may improve over time, I am concerned about how exposure to such stories shapes children’s story quality standards. I am. When these quality indicators are taken away, children’s diverse abilities to understand the meaning of stories are diminished. Despite claims by creators of digital story-writing tools to democratize access to story-making, poorly designed digital books do not allow for stories created digitally and by professional authors. This can unintentionally widen the gap between you and your story. These disparities create an even clearer divide in how literary critics judge what quality literature is worth exposing to children, as opposed to reading materials generated on-demand by AI tools. I am. While the latter may be fun, the former is educational.
Concerns about AI-powered reading coaches, assistants, and tutors pertain to both reading learning and learning. and Read on to learn, especially when it comes to AI-generated prompts.Many digital book production companies have already integrated real-time conversation prompts It has been shown to increase children’s understanding and support literacy development. New AI-generated prompts may be helpful for kids, but not as much as reading with a trained human adult like a teacher, parent, or tutor. Nor should it be used to replace that experience. Overall, these tools have potential, but they can also make things worse. Existing digital divideThis is especially true for children who don’t have access to technology or who don’t have qualified adults to work with them to use technology effectively.
How is research on these tools being developed?
Because the tool is still under development, researchers can only predict rather than judge its effectiveness. Based on academic research on reading motivation, several challenges are anticipated. for example, research result Extrinsic motivators, such as badges, are negatively or only slightly related to reading comprehension. On the other hand, intrinsic reading motivation, which arises from readers’ curiosity and active participation in the reading process, is moderately and positively correlated with measures of reading ability.
In contrast to these findings, AI-powered reading coaches appear to be designed to prioritize promoting extrinsic motivation. Depending on children’s progress and time on the platform, they are rewarded with stickers, applause, and unlockable rewards. Comprehension checks with quizzes can be easily circumvented through trial and error, resulting in children pretending to read and receiving rewards for incorrect answers. Furthermore, the lack of external evaluation to assess whether skills transfer to other texts weakens the accountability of these technologies.
Recent meta-analysis Research into interventions to promote reading motivation has revealed small but notable effects of strategies that customize texts for different reading levels and incorporate connections to the real world. Importantly, this short-term effect is more pronounced among advanced readers than among struggling readers. However, currently, the AI-powered reading coaches on the market lack the specificity of an effective, targeted approach.
It is disappointing to observe these trends. These tools have the potential to improve children’s reading experiences, especially if they are designed with insights from educators and researchers in the field of learning science. For example, these tools can disrupt traditional ideologies of literary texts if they involve teachers in the design process. Through this collaborative approach, teachers’ AI literacy can also be developed. Product developers can obtain information such as: learning science research We build tools that foster self-expression and creativity in children.
Unfortunately, there is a surprising lack of collaboration between the community of edtech companies developing technology products for children and educators and researchers with domain-specific knowledge. Even when companies do engage with researchers, it tends to be sporadic communications and advice rather than ongoing dialogue. And while some companies test tools with teachers, they typically develop features that align with popular features or pressing curriculum requirements rather than the latest and greatest science.
Who suffers most from poor technology? Children. So how can you preserve and encourage learner agency, will, and the ability to make free choices when interacting with an AI-powered reading coach?
Today, this important question boils down to concerns about data privacy and improving data consent collection procedures. However, answering the question also includes determining who ultimately benefits from these tools. If children are the intended beneficiaries, companies building these tools will need to rethink their design and expansion strategies. Rather than rapidly expanding and integrating into various reading products driven by technology trends and investor demands for growth, edtech development requires a more patient approach. This involves participatory design with diverse groups of children and engaging educators and researchers in iterative co-creation cycles. Let’s not diminish the potential of these technologies by hastily releasing tools that are not yet mature enough to fully support children’s development.