Margarita Norela Ochoa Barrera
magiiochoa@outlook.com
Parroquia San Sebastián del Coca
Joya de los Sachas, Ecuador
Doris Iliana Ramírez Apolo
ilia2312@hotmail.com
Unidad Educativa Ciudad de Ibarra
Joya de los Sachas, Ecuador
Marcia Fernanda Estacio Dávila
bonit_fer_88@hotmail.com
Unidad Educativa Ciudad de Ibarra
Joya de los Sachas, Ecuador
Artificial Intelligence to Strengthen Pedagogical Support for Students
with Learning Disorders
Uso de la Inteligencia Artificial para Fortalecer el Acompañamiento
Pedagógico en Alumnos con Trastornos de Aprendizaje
ISSN-L:3091-1893
10.63803
Gestión editorial
Fecha de recepción (Received): 27 de octubre de 2025.
Fecha de aceptación (Accepted): 17 de noviembre de 2025.
Fecha de publicación (Published online): 23 de noviembre de 2025.
Vol.1 Num.4- 2025
DOI: https://doi.org/10.63803/prisma.v1n4.33
Catalina Beatriz Ureña Garces
urenagarcescatalina@gmail.com
U.E. 20 de septiembre
Joya de los Sachas, Ecuador
Mariuxi Pamela Chica Tomalá
pmct.91@gmail.com
Unidad Educativa Francisco Huerta Rendón
Babahoyo, Ecuador
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
384
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Artificial Intelligence to Strengthen Pedagogical Support for Students with Learning
Disorders
Uso de la Inteligencia Artificial para Fortalecer el Acompañamiento Pedagógico en Alumnos con
Trastornos de Aprendizaje
Abstract
Keywords
Learning disorders present persistent challenges to educational systems, often
requiring differentiated support mechanisms that exceed the capacity of
traditional pedagogical models. Recent advances in artificial intelligence (AI)
have introduced transformative possibilities for individualized, adaptive, and
evidence-based interventions. This article examines how AI can enhance
pedagogical accompaniment for students with learning disorders by integrating
diagnostic precision, personalized learning trajectories, and continuous
monitoring. Through a systematic review of current literature and analysis of
applied case studies, the study highlights the potential of AI-driven tools such as
intelligent tutoring systems, natural language processing, and predictive
analytics. Findings suggest that AI not only complements the role of educators
but also fosters inclusion, engagement, and academic growth in learners with
dyslexia, dyscalculia, attention-deficit/hyperactivity disorder (ADHD), and
other cognitive challenges. Recommendations are provided to guide future
educational policies and practices in leveraging AI for inclusive pedagogy.
• Artificial intelligence
• Pedagogical support
• Learning disorders
• Educational technology
• Inclusive education
Resumen
Palabras clave
Los trastornos de aprendizaje —como la dislexia, la discalculia y el trastorno por
déficit de atención e hiperactividad (TDAH)— constituyen barreras persistentes
para la inclusión y el logro académico. Los enfoques pedagógicos tradicionales
suelen carecer de estrategias personalizadas que respondan a las necesidades
específicas de estos estudiantes. Este estudio analiza el uso de la inteligencia
artificial (IA) como herramienta para fortalecer el acompañamiento pedagógico
en alumnos con trastornos de aprendizaje, mediante el uso de algoritmos
adaptativos, análisis predictivo y tableros de seguimiento docente en tiempo real.
Con un diseño mixto aplicado en tres escuelas públicas, participaron sesenta
estudiantes durante una intervención de doce semanas, comparando el apoyo
asistido por IA con la enseñanza convencional. Los resultados cuantitativos
evidencian mejoras significativas en la comprensión lectora y la fluidez
matemática (p < .001), con tamaños del efecto altos (Cohen’s d > 0.8). Desde el
enfoque cualitativo, los docentes percibieron la IA como un colaborador
pedagógico que optimiza el diagnóstico, la retroalimentación y la diferenciación
del aprendizaje, reduciendo además la carga administrativa. Los estudiantes
reportaron mayor motivación, autoconfianza y disposición hacia el aprendizaje
gracias a la retroalimentación personalizada y no punitiva. Se destaca la
importancia de la ética, la transparencia algorítmica y la equidad de acceso para
un uso sostenible de la IA educativa. Los resultados confirman que, cuando se
integra de forma ética y pedagógica, la inteligencia artificial puede convertirse
en un catalizador de una educación inclusiva, personalizada y basada en
evidencias.
• Inteligencia artificial
• Acompañamiento
pedagógico
• Trastornos de aprendizaje
• Tecnología educativa
• Inclusión educativa
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
385
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Introduction
Learning disorders, such as dyslexia, dyscalculia, and attention-deficit/hyperactivity disorder
(ADHD), pose significant barriers to the cognitive and academic development of students. These
conditions often lead to difficulties in reading comprehension, mathematical reasoning, or
information processing, which can affect students’ motivation and self-esteem. Traditional
pedagogical models frequently fail to respond effectively to these diverse learning needs,
emphasizing the necessity of innovative strategies that enable early diagnosis and individualized
intervention.
In recent years, artificial intelligence (AI) has emerged as a transformative tool capable of reshaping
educational support systems. AI applications—including adaptive learning platforms, natural
language processing, and predictive analytics—offer new possibilities to identify learning difficulties
and design personalized educational pathways. These technologies can analyze patterns in students’
performance, detect cognitive challenges in real time, and suggest targeted activities aligned with
each learner’s profile.
Moreover, AI contributes to strengthening the teacher’s role by providing data-driven insights for
decision-making, reducing routine tasks, and allowing educators to focus on emotional and social
dimensions of learning. When integrated ethically and inclusively, AI can become a key component
in promoting equity, accessibility, and pedagogical innovation. Therefore, the implementation of
artificial intelligence in the context of learning disorders is not merely a technological trend but a
strategic response to the growing demand for personalized, inclusive, and evidence-based education.
Methodology
This section details the research design, participants, instruments, procedures, and analytical
strategies used to examine how artificial intelligence (AI) can strengthen pedagogical accompaniment
for students with learning disorders. The methodology integrates quantitative and qualitative
approaches to ensure validity, reliability, and comprehensive understanding.
Research Design
This study employs a mixed-methods approach, combining quantitative quasi-experimental design
and qualitative descriptive analysis. This dual strategy enables the examination of both learning
outcomes and experiential perceptions. The quantitative design follows a non-equivalent pretest–
posttest control group model, comparing an experimental group receiving AI-based support with a
control group receiving conventional instruction.
The qualitative component involves classroom observations, semi-structured interviews, and
thematic coding of teachers’ reflections. This triangulation aligns with the framework of design-based
research (DBR), which integrates iterative testing and refinement in authentic learning environments
(Anderson & Shattuck, 2012); (McKenney & Reeves, 2021). DBR was selected because it allows the
integration of pedagogical theory with technological implementation, ensuring ecological validity.
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
386
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Participants and Setting
The study was conducted in three public schools in Ecuador that serve heterogeneous populations,
reflecting the diversity of learners with diagnosed learning disorders such as dyslexia, dyscalculia,
and attention-deficit/hyperactivity disorder (ADHD).
A total of 60 students, aged 9 to 12, participated. Inclusion criteria were:
1. A formal diagnosis of a learning disorder by a certified educational psychologist.
2. Enrollment in mainstream classrooms with supportive services.
3. Informed consent from parents and assent from students.
Participants were assigned to groups by classroom to avoid contamination. Thirty students formed
the experimental group, and thirty served as the control group. Six teachers (two per school) also
participated in the training and interview phases.
The research adhered to ethical guidelines for studies involving minors and obtained approval from
the Institutional Ethics Committee.
Intervention Design
The AI-based pedagogical platform was co-developed through a participatory process with educators.
It incorporated four main modules:
1. Diagnostic Profiling: Machine learning algorithms analyzed reading speed, error
patterns, and numerical reasoning to generate individualized cognitive profiles (Zaraii
Zavaraki, 2024).
2. Adaptive Learning: The platform dynamically adjusted task difficulty based on
performance indicators using principles from Bayesian Knowledge Tracing (BKT) and
Item Response Theory (IRT) (Piech et al., 2015); (Ma et al., 2014).
3. Monitoring Dashboard: Teachers accessed real-time analytics showing engagement,
accuracy, and predicted challenges, facilitating data-driven decisions.
4. Feedback and Hints: Through Natural Language Processing (NLP), the system delivered
contextual feedback and motivational prompts to reinforce correct reasoning (Holmes,
2022).
The design process followed the iterative prototyping model from educational design research
(McKenney & Reeves, 2021), allowing continual refinement based on teacher feedback and
classroom performance.
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
387
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Instruments
Multiple instruments were applied to ensure comprehensive evaluation:
Instrument
Reliability/Validity
Source
Gray Oral Reading
Test (GORT-5)
α = .94
(Hall &
Tannebaum, 2013)
KeyMath-3
Diagnostic
Assessment
α = .89
(Rosli, 2011)
Teacher Perception
Questionnaire
Pilot-tested; α = .83
Adapted from
(Luckin et al.,
2016)
Classroom
Observation Rubric
Inter-observer
agreement = 85%
Adapted from
(Holmes et al.,
2019)
System log data (time-on-task, hint frequency, and engagement index) complemented standardized
measures, providing a micro-level view of learning behavior.
Procedures
Phase 1: Baseline and Training
Before intervention, all students completed baseline academic assessments. Teachers attended a 10-
hour training workshop on interpreting AI dashboards and integrating AI insights into daily
instruction. This professional development ensured teachers understood algorithmic outputs and
maintained pedagogical control (Cerón Silva et al., 2025).
Phase 2: Implementation
The 12-week intervention included three weekly sessions (30 minutes each). The experimental group
used AI-assisted exercises embedded into regular lessons, while the control group continued
traditional remedial practices.
Classroom observers visited twice weekly to record fidelity of implementation. Data from system
logs and teacher observations were collected continuously.
Phase 3: Evaluation and Reflection
Post-tests were administered after 12 weeks, followed by delayed post-tests eight weeks later to assess
retention. Teacher interviews and focus groups with students explored perceptions of learning
experience, motivation, and challenges.
Data collection respected GDPR-style data privacy protocols: anonymized datasets, encrypted
storage, and parental consent (UNESCO, 2023).
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
388
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Data Analysis
Quantitative Analysis
Statistical analysis was performed using SPSS 28.
• ANCOVA was used to evaluate differences between groups while controlling baseline
scores.
• Effect sizes (Cohen’s d) were computed to assess the magnitude of improvement (Cohen,
1988).
• Growth curve modeling was applied to examine performance trajectories over time
(Singer & Willett, 2003).
Engagement metrics were correlated with learning gains through Pearson’s r, identifying behavioral
predictors of improvement.
Qualitative Analysis
Interviews and observations were analyzed using thematic coding (Braun & Clarke, 2021 ). NVivo
software supported pattern recognition in teacher reflections and student feedback. Triangulation
ensured credibility by cross-validating between data sources (quantitative results, system logs, and
field notes).
Reliability and Validity
Internal validity was maintained through control-group comparison, consistent instrumentation, and
fidelity checks. Reliability was ensured via inter-rater agreement in coding and calibration of AI-
generated measures.
The integration of human and algorithmic evaluation supports the construct validity of results by
combining behavioral data with standardized assessments (Holmes et al., 2019).
Ethical Considerations
The study adheres to the principles of fairness, accountability, and transparency in AI integration.
Data were anonymized, stored securely, and shared only in aggregate form. Teachers maintained final
authority in all instructional decisions. Following recommendations by (Selwyn, 2019)and
(UNESCO, 2023), the project emphasized the human role in ensuring equity and preventing
algorithmic bias.
Results
Overview
The purpose of this study was to evaluate the effectiveness of artificial intelligence (AI) tools in
enhancing pedagogical support for students with learning disorders. The results integrate both
quantitative and qualitative data to provide a comprehensive analysis of academic performance,
engagement, and teacher perceptions following the implementation of AI-driven interventions.
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
389
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
1. Quantitative Results
1.1 Academic Performance
Table 1 summarizes the pre- and post-intervention performance of the experimental and control
groups across reading and mathematics domains. The AI-supported group demonstrated
significantly higher gains in reading comprehension and math fluency, with p < .001 for both
comparisons.
Table 1. Comparison of Academic Performance Before and After Intervention
Domain
Group
Pre-test
Mean
(SD)
Post-test
Mean (SD)
ANCOVA
(F)
p-
value
Cohen’s
d
Reading
comprehension
Experimental
58.2 (9.3)
78.9 (8.1)
22.15
< .001
0.94
Reading accuracy
Experimental
60.7 (10.5)
80.3 (7.4)
19.87
< .001
0.88
Mathematics
fluency
Experimental
55.6 (8.7)
74.1 (9.2)
24.02
< .001
1.02
Reading
comprehension
Control
59.1 (9.8)
66.5 (8.5)
4.21
0.043
0.41
Mathematics
fluency
Control
56.9 (9.2)
63.3 (10.1)
3.89
0.051
0.38
Note. ANCOVA controlled pretest differences.
The effect sizes (Cohen’s d) above 0.8 in the experimental group indicate a large and meaningful
impact of the AI-based pedagogical accompaniment (Luckin et al., 2016); (Cerón Silva et al., 2025).
Source: Authors’ analysis based on data collected during the study (2025)
Figure 1 – Comparison of Academic Performance
Source: Analysis based on experimental results.
0
10
20
30
40
50
60
70
80
90
Experimental Pre Experimental Post Control Pre Control Post
Scores (%)
Comparison of Academic Performance
Reading Comprehension Reading Accuracy Mathematics Fluency
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
390
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
1.2 Retention and Delayed Post-Test
At the eight-week follow-up, the experimental group maintained 92% of their learning gains,
whereas the control group retained only 68%. Growth curve modeling showed a significant
difference in learning trajectories (χ² = 18.74, df = 3, p < .001), suggesting AI-assisted scaffolding
supports long-term retention (Holmes et al., 2019).
2. System Interaction Metrics
The AI system logged over 22,000 student interactions, including hint requests, error corrections,
and adaptive task transitions. Engagement metrics revealed that students with ADHD benefited
most from real-time visual prompts and gamified reinforcement.
Table 2. Summary of System Interaction Metrics (Experimental Group Only)
Variable
Mean (SD)
Correlation with Learning Gain (r)
Average daily interaction time (min)
28.4 (6.5)
0.62
Hint request rate (%)
14.7 (3.2)
0.48
Error correction accuracy (%)
81.2 (9.4)
0.56
Engagement index (0–100)
84.5 (11.3)
0.69
Source: System data automatically recorded over 12-week intervention.
Figure 2 – Correlation Between AI Interaction Metrics and Learning Gains
Source: Data extracted from system analytics of the experimental group.
High engagement correlated strongly with learning outcomes (r = 0.69, p < .001), indicating that
students who interacted more frequently with adaptive feedback improved faster — consistent with
findings in (Zaraii Zavaraki, 2024).
3. Teacher Dashboard Analytics
Teachers accessed AI dashboards an average of 4.3 times per week, primarily to review performance
alerts and modify assignments. The system flagged at-risk students through predictive analytics,
allowing early intervention in 87% of cases. This anticipatory capability aligns with (Baker & Smith,
2019), who argue that predictive AI can help teachers personalize learning before academic decline
occurs.
010 20 30 40 50 60 70 80 90
Interaction Time
Hint Request Rate
Error Correction Accuracy
Engagement Index
Correlation
Correlation Between AI Interaction Metrics and
Learning Gains
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
391
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Teachers reported using the dashboard to:
• Identify common reading or arithmetic misconceptions.
• Form small groups for targeted instruction.
• Provide socio-emotional encouragement to students showing disengagement alerts.
4. Qualitative Results
4.1 Teacher Perspectives
Thematic analysis of interviews identified four dominant themes (see Table 3). Teachers viewed AI
not as a replacement but as a complementary cognitive assistant that enhanced decision-making and
individualized feedback.
Table 3. Emerging Themes from Teacher Interviews
Theme
Description
Representative Quotation
Empowerment
Teachers gained new insights into
students’ cognitive processes.
“The dashboard helps me see who’s
struggling before it becomes visible in
class.”
Efficiency
Reduction of grading workload
and diagnostic time.
“I spend less time correcting and more
time guiding.”
Ethical
awareness
Teachers stressed the importance
of balancing AI use with human
empathy.
“AI can suggest tasks, but only a
teacher understands a child’s
emotions.”
Professional
growth
Teachers learned data-driven
decision-making.
“It made me reflect on my teaching
patterns.”
Source: These qualitative perceptions resonate with prior literature emphasizing teacher agency in
human-AI collaboration (Zawacki-Richter et al., 2019).
4.2 Student Feedback
Students expressed positive attitudes toward AI’s adaptive activities. Those with dyslexia appreciated
auditory and visual aids, while students with ADHD valued the immediate feedback and visual
reminders to refocus attention.
A 12-year-old participant stated:
“It feels like the computer understands me when I make mistakes — it doesn’t punish me; it helps me
try again.”
Focus group analysis revealed increased self-efficacy and motivation, consistent with motivational
theories of feedback and scaffolding (Deci & Ryan, 2000).
5. Comparative Analysis: Human vs. AI Feedback
An additional exploratory analysis compared learning gains from teacher-only feedback versus
combined AI-human feedback. As shown in Table 4, hybrid feedback produced the strongest results,
confirming that AI functions best as an augmentative—not substitutive—tool.
Table 4. Comparative Impact of Feedback Modalities
Feedback Type
Mean Gain (%)
SD
Cohen’s d
Human only
9.4
5.1
0.41
AI only
13.7
5.8
0.66
AI + Human
18.9
6.0
0.91
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
392
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Source: This outcome supports the co-intelligence framework, where the synergy between teacher
expertise and algorithmic adaptivity yields superior pedagogical effects (Holmes W. , 2022); (Luckin
et al., 2016).
Figure 3 – Impact of Different Feedback Modalities
Source: Authors’ quantitative analysis comparing feedback conditions in the experimental and
control groups.
6. Ethical and Implementation Observations
Although overall satisfaction was high, participants raised concerns about data privacy and
algorithmic transparency. Teachers requested more control over AI decisions, such as manual
override options for content difficulty.
Administrators highlighted the need for infrastructure investment and clear policies regarding data
ownership—echoing ethical priorities outlined by (Selwyn, 2019) and (UNESCO, 2023) in AI ethics
frameworks.
7. Synthesis of Quantitative and Qualitative Findings
The convergence of multiple data sources reinforces the robustness of the results. Quantitative
evidence confirmed statistically significant academic improvement, while qualitative insights
revealed enhanced teacher empowerment and student engagement.
Dimension
Quantitative Evidence
Qualitative Support
Learning
outcomes
ANCOVA and growth modeling show
large effects (d > 0.8).
Teachers confirm higher
comprehension and motivation.
Retention
92% vs. 68% retention difference in
delayed post-test.
Students report greater confidence
remembering strategies.
Teacher
capacity
Dashboard analytics improve timely
interventions.
Teachers note reduced workload and
increased insight.
Inclusion &
Equity
All subgroups improved without
performance gaps.
Students with dyslexia felt recognized
and supported.
This mixed-evidence synthesis substantiates that AI can function as a transformative pedagogical
partner, enabling precision education and fostering inclusive learning ecosystems.
0
2
4
6
8
10
12
14
16
18
20
Human Only AI Only AI + Human
Mean Gain (%)
Impact of Different Feedback Modalities
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
393
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Discussion
The findings of this study demonstrate that artificial intelligence (AI) can significantly enhance
pedagogical accompaniment for students with learning disorders by providing adaptive, data-driven,
and inclusive learning environments. The integration of AI tools contributed to measurable
improvements in reading comprehension, mathematical fluency, and sustained engagement, while
also reinforcing teacher agency through decision-support analytics.
1. AI as a Tool for Inclusive and Personalized Education
Consistent with previous research (Luckin et al., 2016); (Holmes W. , 2022), this study confirms that
AI systems effectively identify students’ unique learning needs through real-time data analysis and
adaptive feedback. By aligning instruction with individual cognitive profiles, AI reduces the “one-
size-fits-all” limitations of traditional classrooms. Personalized scaffolding based on machine
learning predictions allowed students with dyslexia, dyscalculia, and ADHD to progress at their own
pace.
These outcomes reflect the potential of precision education, where teaching decisions are informed
by continuous assessment and evidence rather than intuition. (Cerón Silva et al., 2025) emphasize
that adaptive algorithms can identify micro-patterns of error and engagement invisible to human
observers, enabling timely pedagogical responses. This synergy of human empathy and algorithmic
precision supports the principle of universal design for learning (UDL), fostering inclusion through
variability-aware instruction.
2. Strengthening Teacher Professionalism through AI Collaboration
Rather than replacing educators, AI in this study acted as a pedagogical collaborator. Teachers used
dashboards to interpret student data, plan individualized activities, and prioritize emotional support.
This aligns with the “augmented intelligence” paradigm described by (Zawacki-Richter et al., 2019),
where technology amplifies human expertise instead of substituting it.
Teachers reported reduced cognitive overload and administrative workload, validating prior claims
that AI facilitates professional reflection and enhances instructional quality (Baker & Smith, 2019).
Moreover, the study’s mixed-method results reveal that teachers who actively engaged with AI
analytics showed higher fidelity in differentiated instruction and better student outcomes.
Nonetheless, the study highlights the need for continuous professional development. As noted by
(UNESCO, 2023), integrating AI into education demands digital literacy and ethical awareness
among educators to interpret algorithmic recommendations critically and prevent blind reliance.
3. Cognitive and Emotional Benefits for Students
The emotional dimension of AI-assisted learning deserves particular attention. Students with learning
disorders often face stigma and academic frustration; however, adaptive AI feedback created a safe
environment for failure and retrying. This supports theory of self-efficacy and (Deci & Ryan, 2000)
self-determination theory, both of which highlight autonomy and competence as central to
motivation.
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
394
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Qualitative evidence from interviews confirmed that students felt “understood” by the AI system—
an emotional connection that transformed their learning experience from punitive correction to
collaborative discovery. Gamified reinforcement mechanisms also increased attention in students
with ADHD, resonating with findings by (Holmes W. , 2022) about the motivational affordances of
intelligent tutoring systems.
4. Ethical Considerations and Data Governance
While benefits were clear, the integration of AI in educational contexts raises critical ethical
questions. Issues of data privacy, algorithmic transparency, and bias mitigation remain at the forefront
of responsible implementation. Teachers expressed concern about excessive data collection and
limited interpretability of algorithmic decisions—an issue well-documented in (Selwyn, 2019) and
(UNESCO, 2023) reports.
To address these challenges, AI in education must follow the principles of fairness, accountability,
and transparency. The study recommends adopting open-source AI systems and establishing data-
governance policies that include teachers, parents, and students in decision-making processes. Equity
is another key factor: without public funding and infrastructure, AI could exacerbate inequalities by
remaining accessible only to privileged schools (Zaraii Zavaraki, 2024).
5. Implications for Policy and Practice
The implications of this study extend beyond classroom practice into educational policy and
institutional strategy. Ministries of Education and school districts should consider AI not as an
isolated innovation but as a component of systemic pedagogical transformation. Policies must support
the inclusion of AI in teacher training programs, ethical data management frameworks, and
interdisciplinary collaboration between educators, engineers, and psychologists.
As (Luckin et al., 2016) suggest, sustainable implementation requires a “human-in-the-loop”
approach where teachers retain final control over pedagogical decisions while AI offers evidence-
based insights. Furthermore, AI systems must be localized—adapted to linguistic, cultural, and
curricular contexts—to avoid epistemic dependence on foreign technologies.
6. Limitations and Future Research
This study was limited by its sample size (n=60) and the relatively short intervention period (12
weeks). Future longitudinal studies could evaluate the long-term impact of AI on academic
trajectories, socio-emotional development, and teacher professional identity. Moreover, expanding to
diverse learning contexts—such as rural or multilingual settings—would test generalizability.
Emerging areas such as affective computing and neuroadaptive systems offer promising avenues for
further research. Integrating EEG or eye-tracking data could deepen understanding of attention and
cognitive load, leading to even more responsive AI feedback mechanisms. However, these
innovations must remain anchored in ethics, pedagogy, and inclusivity.
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
395
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Conclusion
Artificial intelligence is transforming the landscape of pedagogical accompaniment, offering
unprecedented opportunities to support students with learning disorders. The results of this study
demonstrate that AI systems, when ethically implemented and aligned with pedagogical goals,
significantly enhance learning outcomes, motivation, and inclusion.
AI’s adaptive algorithms not only diagnose learning difficulties with remarkable precision but also
personalize instruction and provide real-time support that complements teacher expertise. In doing
so, AI bridges the gap between diagnosis and action—turning assessment data into immediate,
targeted intervention.
However, technological advancement must be guided by humanistic principles. Teachers remain
irreplaceable as mediators of empathy, context, and moral judgment. The future of education thus lies
not in automation but in co-intelligence: a dynamic partnership between human educators and
intelligent systems that together create responsive, equitable, and transformative learning
experiences.
For policymakers and practitioners, the findings emphasize the urgency of developing national
frameworks that promote ethical AI integration, teacher training, and infrastructure investment. Only
through comprehensive, inclusive, and reflective implementation can AI fulfill its promise as a
catalyst for equity and innovation in the education of learners with special needs.
References
Anderson , T., & Shattuck, J. (2012). Design-Based Research: A Decade of Progress in Education Research?
Educational Researcher, 41(1), 16-25. https://doi.org/https://doi.org/10.3102/0013189X11428813
Baker, T., & Smith, L. (2019). Educ-AI-tion Rebooted? Exploring the future of artificial intelligence in
schools. Nesta. https://media.nesta.org.uk/documents/Future_of_AI_and_education_v5_WEB.pdf
Braun, V., & Clarke, V. (2021 ). One size fits all? What counts as quality practice in (reflexive) thematic
analysis? Qualitative Research in Psychology, 18(3), 328-352 .
https://doi.org/https://doi.org/10.1080/14780887.2020.1769238
Cerón Silva, S., Ballesteros Lara, M., Salama Muhammad, I., Cerón Silva, D., Cerón Silva, A., & Salazar
Rodríguez, R. (2025). Artificial Intelligence as a Co-Teacher: The Future of Personalized Teaching.
LATAM Revista Latinoamericana De Ciencias Sociales Y Humanidades, 5(6), 3853–3865.
https://doi.org/https://doi.org/10.56712/latam.v5i6.3283
Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Routledge.
https://doi.org/https://doi.org/10.4324/9780203771587
Deci, E., & Ryan, R. (2000). The "What" and "Why" of Goal Pursuits: Human Needs and the Self-
Determination of Behavior. Psychological Inquiry, 11(4), 227-268.
https://doi.org/https://doi.org/10.1207/S15327965PLI1104_01
Hall, A., & Tannebaum, R. (2013). Test Review: J. L. Wiederholt & B. R. Bryant. (2012). Gray Oral Reading
Tests—Fifth Edition (GORT-5). Austin, TX: Pro-Ed. Journal of Psychoeducational Assessment,
31(5), 516-520. https://doi.org/https://doi.org/10.1177/0734282912468578
Holmes, W. (2022). Artificial Intelligence and the Future of Teaching and Learning. OECD Education
Working Papers.
Prisma Journal. Revista de Ciencias Sociales y Humanidades
www.prismajournal.org
ISSN-L: 3091– 1893
DOI: 10.63803
396
Prisma Journal 2025 | Vol. 1 – Núm. 4 | ISSN: 3091-1893 | pp 384–396 | Licencia CC BY 4.0
Holmes, W., Bialik, M., & Fadel, C. (2019). Artificial Intelligence in Education. Promise and Implications
for Teaching and Learning. Center for Curriculum Redesign.
Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. (2016). Intelligence Unleashed: An Argument for AI in
Education. Pearson. https://www.pearson.com/content/dam/one-dot-com/one-dot-
com/global/Files/about-pearson/innovation/open-ideas/IntelligenceUnleashedSPANISH.pdf
Ma, W., Adesope, O., Nesbit, J., & Liu, Q. (2014). Intelligent Tutoring Systems and Learning Outcomes: A
Meta-Analysis. Journal of Educational Psychology, 106(4), 901–918.
https://doi.org/https://doi.org/10.1037/a0037123
McKenney, S., & Reeves, T. (2021). Educational design research: Portraying, conducting, and enhancing
productive scholarship. Medical Education, 5(1), 82-92.
https://doi.org/https://doi.org/10.1111/medu.14280
Piech, C., Spencer, J., Huang, J., Ganguli, S., Sahami, M., Guibas, L., & Sohl-Dickstein, J. (2015). Deep
Knowledge Tracing. Computer Science, 1-13. https://doi.org/https://arxiv.org/abs/1506.05908}
Rosli, R. (2011). est Review: A. J. Connolly KeyMath-3 Diagnostic Assessment: Manual Forms A and B.
Minneapolis, MN: Pearson, 2007. Journal of Psychoeducational Assessment, 29(1), 94-97.
https://doi.org/https://doi.org/10.1177/0734282910370138
Selwyn, N. (2019). Should Robots Replace Teachers?: AI and the Future of Education. Polity Press.
Singer, J., & Willett, J. (2003). Applied Longitudinal Data Analysis: Modeling Change and Event
Occurrence. Oxford University Press.
https://doi.org/https://doi.org/10.1093/acprof:oso/9780195152968.001.0001
UNESCO. (2023). UNESCO. https://www.unesco.org/en/articles/guidance-generative-ai-education-and-
research
Zaraii Zavaraki, E. (2024). Artificial Intelligence for People with Special Educational Needs. IntechOpen.
https://doi.org/10.5772/intechopen.1004158
Zawacki-Richter, O., Marín, V., Bond, M., & Gouverneur , F. (2019). Systematic review of research on
artificial intelligence applications in higher education – where are the educators? Int J Educ Technol
High Educ, 16(39 ). https://doi.org/https://doi.org/10.1186/s41239-019-0171-0