Artificial intelligence (AI) is gradually infiltrating the world of education, particularly in the teaching of reading. More and more digital tools are offering to help budding readers decipher their first texts.

While official data shows large gaps in achievement between students and significant difficulties for some students entering sixth grade, what can we expect from these new technological supports? Are they truly effective? Can they complement traditional methods? Are they likely to replace them, or is this just a science fiction scenario?

From decoding to understanding: a long learning process

Let us remember that reading is defined as “the action of deciphering a text, identifying its characters and words in order to understand its meaning” . This requires rigorous learning which involves “developing skills in two areas: the identification of written words and the processing of meaning, for the understanding of texts” .

Before you can understand a text, it’s crucial to master decoding. This means matching what you see (graphemes) with what you hear (phonemes). For example, when a child sees the word “boat,” they must first identify the graphemes b, a, t, and water, then connect them to the corresponding sounds /b/, /a/, /t/, and /o/. By combining these sounds, they can recognize the word “boat,” especially if they’ve heard it around them before.

This learning begins with simple and common rules, which become more complex but become automatic over time. Then comes comprehension, which is based on several skills (semantic, syntactic, morphological processing) and is not exclusive to reading.

Is learning to read an obstacle course? According to INSEE (2011) , one in five students entering secondary school encounters difficulties with writing. In addition, between 20 and 40% of students of this age have difficulty understanding texts. These difficulties unfortunately do not diminish over time. In adolescence, one in ten young people have very poor reading comprehension, accompanied by a significant vocabulary deficit.

Reading skills therefore do not develop uniformly among all students and can be influenced by social and cultural factors . In more advantaged environments, reading is often valued as a leisure and escape activity, while in less advantaged environments, it is mainly seen as a learning tool . This difference in usage can have repercussions on students’ motivation and engagement in reading, thus influencing their progress and level of comprehension.

Digital tools: opportunities and challenges

The many digital educational tools developed over the past twenty years can be divided into two broad categories: those focused on decoding and those aimed at comprehension. Among the tools dedicated to decoding, we can cite Daisy Quest and Daisy Castle , Lexia , Play-On and GraphoGame . Regarding the development of reading comprehension, several resources have also emerged, such as ELAN , ABRACADABRA and ePearl .

These tools each pursue specific learning objectives, such as the development of phonological skills, mastery of grapheme-phoneme correspondences, improvement of reading fluency, and text comprehension. In addition, they target diverse populations based on educational needs.

However, to ensure their effectiveness, it is essential that they are designed in close connection with scientific research. This allows not only to verify their reliability objectively (for example GraphoGame ), but also to precisely evaluate their specific effects .

While these tools have been widely adopted by primary school teachers ( 94% in 2015 ) for lesson preparation, their integration in the classroom remains limited. According to the international TALIS 2018 survey , only 14.5% of primary school teachers report frequently or systematically allowing Information and Communication Technologies (ICT) when students are carrying out projects or assignments in class.

From a student perspective, these tools increase motivation and attention, making learning more engaging. Immediate feedback promotes understanding and progress, while the individualization of the learning path allows for tailoring learning to each student’s specific needs.

That said, these technologies are not without limitations  : cognitive overload, sleep disruption, unequal access, and even risks of social isolation. These issues underscore the need for thoughtful and moderate use of digital tools, as well as adequate teacher training to fully exploit their potential in teaching reading.

AI, an educational revolution?

More recently, AI has established itself in all disciplines, renewing the tools available. In France, regarding learning to read, we can cite applications like Lalilo and Navi .

Like traditional digital tools, AI allows for immediate feedback but also goes further by personalizing each learner’s journey . It allows for real-time adjustment of the difficulty of exercises and offers content adapted to the student’s progress, which offers differentiated and more interactive learning .

It is a valuable tool for supporting teachers and reducing repetitive tasks. It facilitates the creation of quizzes and assessments, allowing teachers to save time while targeting students’ difficulties and thus adapting their interventions.

To date, few studies have examined the impact of AI on learning to read. While some research has highlighted its benefits in comprehension, story structuring, and vocabulary enrichment, its effectiveness in the early stages of learning remains to be demonstrated.

Although AI is transforming many fields, its integration into education still raises questions and challenges. Its effectiveness must still be compared to traditional methods to assess its true benefits. Furthermore, the use of adaptive and generative tools requires technical skills that teachers do not always possess, making specific training essential.

Finally, on a cognitive level, AI, although interactive and engaging, could cause cognitive overload , particularly in young children, who must juggle between the digital interface and fundamental learning.

Overreliance on these tools could also limit their autonomy and their ability to develop effective reading strategies. Thus, while promising, AI in education must be used with moderation and discernment to maximize its benefits without suffering negative effects. More broadly, its financial impact, as well as its ecological impact, in terms of energy consumption, remains to be determined and will need to be taken into account in future research.

Author Bios: Xavier Aparicio is University Professor of Cognitive Psychology, Camille Pistre is a Doctoral student and Ugo Ballenghein is University Professor – Cognitive Psychology all at the University of Paris-Est Créteil Val de Marne (UPEC)