* For more information regarding media usage, ownership and rights please contact
Art × Science International.
PARIS, France – May 19, 2025 – Non-profit innovator Art × Science International today
detailed the advanced AI technology underpinning its mission to adapt literary classics for
children and adults with ADHD. By integrating sophisticated Natural Language Processing
(NLP) using GPT-like architectures, machine learning models, and AI-driven content
generation, guided by educational neuroscience experts, Art × Science is creating uniquely
accessible reading experiences while preserving the soul of classic literature. All profits are
reinvested into expanding this high-tech, accessible content in France.
The organization's core challenge is to make dense literary works digestible and engaging
for minds prone to distraction, without sacrificing the original author's unique style or
narrative depth. This is achieved through a proprietary AI platform that performs multi-level
textual analysis and adaptation. This includes:
● Advanced NLP for stylometric preservation & complexity reduction: Art ×
Science employs cutting-edge NLP algorithms in addition to models inspired by
GPT-like architectures and fine-tuned on literary corpora. These models analyze
syntactic structures, semantic nuances, and author-specific stylistic markers
(stylometry). This allows the AI to identify overly complex sentences and dense
paragraphs, which are then intelligently rephrased or broken down. Crucially, the AI
is trained to retain the original authorial voice, ensuring stylistic integrity even as
readability is enhanced for ADHD individuals.
● AI-Driven chapter segmentation & narrative pathing: To combat cognitive
overload, the AI platform automatically identifies optimal breakpoints for shorter
chapters. For its pioneering non-linear reading feature, the system constructs "story
graphs" using graph databases on Neo4j, mapping characters, events, themes,
relationships, and crucially, the sequential "plot steps" or elements of the
narrative. These graphs are visualized, providing users with a clear
representation of the story's structure. The AI utilizes graph traversal
algorithms, such as depth-first search and breadth-first search, to suggest
alternative pathways through the narrative. This enables readers to follow
specific interests (e.g., a character's arc, a particular subplot) while maintaining
coherence and engagement.
● Intelligent visual support generation: Recognizing the importance of multi-modal
input, Art × Science leverages AI to inform the creation of visual aids. While not
generating final art, the AI analyzes text for descriptive passages and emotional tone
to suggest opportune moments and thematic content for illustrations and visual cues.
This helps to reduce mental imagery load and reinforce comprehension.
● Scalable adaptation engine: The AI is designed for scalability, enabling the
ambitious goal of adapting 4,500 books in three languages. Our initial focus is
comprehensive French content, where we have direct access to educator
feedback. These will then be translated into English and Spanish. This involves
automated initial adaptation passes, which are then reviewed and refined by human
literary and neurodiversity experts, creating an efficient human-AI collaborative
workflow. We currently rely heavily on human data labeling to identify notable
plot points, events, dialogue, and "hard to read" passages to train our models.
Why This Is a Game-Changer for Tech & AI Media:
● Sophisticated NLP in action: Art × Science isn't just using off-the-shelf AI. They
are developing and fine-tuning GPT-like architectures and other NLP models
specifically for the complex task of literary adaptation for neurodivergent individuals,
focusing on the delicate balance between accessibility and artistic fidelity.
● Human-AI collaboration in content creation: This initiative showcases an AI
model augmenting human expertise. Educational neuroscientists and literary
specialists guide the AI's development and curate its output, ensuring pedagogical
soundness and literary quality. Significant human effort is currently dedicated to
labeling data that the AI learns from.
● Building "Story Graphs" for interactive narratives: The technology behind a
non-linear reading path – involving knowledge graph construction from unstructured
text– is at the forefront of interactive storytelling and personalized content delivery.
These are not simple static graphs, but visualized structures of the story
including "plot steps".
● AI for social change: This is a prime example of AI being ethically deployed to
address significant societal challenges, specifically enhancing cognitive accessibility
and promoting inclusivity in cultural engagement.
● The future of personalized education tech: The underlying AI framework has
potential applications beyond literature, hinting at the future of AI-driven personalized
learning tools for diverse cognitive profiles.
"Our AI is not about replacing human creativity but augmenting it to break down barriers to
literature" explained Noé Cabannes Michel, executive at Art × Science International. "We're
essentially teaching our models the art of literary empathy – understanding both the
author's intent and the reader's cognitive needs. The goal is to transform classic texts into
dynamically accessible experiences."
Roadmap & Impact:
● Opening registration for crowdsourcing with microtasks available.
● Further development of the "Sapere aude" project, which will leverage these core AI
technologies.
● Distribution of initial copies to teachers, journalists and early supporters for feedback.
● Progressive launch of 4,500 AI-enhanced titles in book stores.
Testimonials from early trials continue to highlight the positive impact, with parents and
educators noting significant improvements in engagement and comprehension among
children with ADHD.
How to get Involved & technical deep dives:
Art × Science invites technology partners, AI researchers, and individuals to support its
mission. Donations are vital for ongoing R&D and content deployment. For a more detailed
exploration of their technology stack, potential white papers, or to discuss technical
collaboration, contact Noé Cabannes Michel.
For general information, newsletter subscriptions, or donations, please visit
Contact: Noé Cabannes Michel noe@artxsc.org +33636062594 https://artxsc.org
###
High resolution screenshot available with edition and commercial use allowed:
https://artxsc.org/images/press-preview-sapare-aude-gui.png
NB: You can also use “Art x Science International” instead of “Art × Science International” for the
organization’s name.
Distributed by Pressat