Assessing the impact of AI tools on mobility and daily assistance for children with down syndrome in Saudi Arabia

Abstract

This mixed-methods study investigated the impact of AI-powered assistive technology on mobility, communication, and daily living assistance in children with Down syndrome in Saudi Arabia. We looked at information from 123 carers (47 who used AI and 76 who did not) with similar child backgrounds through structured surveys and seven detailed interviews, using careful statistical methods and thorough initial assessments. AI users found it much easier to do things compared to those who didn’t use AI, showing a medium improvement in moving around (Cohen’s d = 0.65) and communicating (d = 0.72) and a large improvement in household tasks (d = 0.83). Among AI users only, perceived effectiveness was rated moderately high for daily independence (M = 3.85, SD = 0.93) and mobility (M = 3.78, SD = 1.18). Seven interviewed carers reported subjective stress reduction when children used AI tools independently. This secondary benefit extends the impact of AI beyond direct child outcomes to broader family well-being. Primary barriers to adoption included financial constraints (74.46%), insufficient technical support (38.29%), and technology adaptation difficulties (31.91%). Cultural perceptions and the absence of Arabic-localised AI solutions pose additional challenges that align with the global patterns identified in cross-cultural research. Although we did not directly test cross-cultural comparisons, this study suggests that Saudi Arabia’s adoption barriers may share similarities with international patterns. Policy recommendations include establishing government subsidies covering 60–80% of AI tool costs, mandating insurance coverage, creating regional AI support centres, and prioritising the development of culturally adapted Arabic language solutions.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-15085-7.

Introduction

In Saudi Arabia, approximately one in every 554 live births results in a child with Down syndrome who faces mobility and daily living challenges that significantly impact both children and their families¹. Many studies have shown that children with Down syndrome experience difficulties with muscle development. Research has demonstrated that problems with postural control make it difficult for them to perform functional activities and participate in daily life². It is clear that mobility issues make it difficult for people to do things on their own. This finding is supported by Souto et al.³, who discovered that environmental, personal, and activity factors strongly affect people’s participation in daily activities. This impact extends beyond individual deficiencies to include family relations and socialisation. In particular, Saudi families face unique challenges in accessing appropriate support services, as indicated by recent studies⁴. As Saudi Vision 2030 aims to promote economic diversification and technological augmentation, attention to the use of creative technologies to enhance the provision of healthcare services within the sector is growing⁵. This presents fresh opportunities for addressing such challenges using artificial intelligence tools, so it provides a hopeful future for improving the quality of life of Saudi Arabian children with Down syndrome.

Involving artificial intelligence assistive technologies has shown positive outcomes in helping people with disabilities, which De Freitas et al.⁶ mention as a transformative change in the use of assistive technologies. This is particularly true in Saudi Arabia. Recent studies were encouraging, and White et al.⁷ achieved substantial gains in independence, living conditions, and adaptive behaviour after applying assistive devices. White et al.⁷ reported notable increases in independence, quality of life, and adaptive behavioural skills following assistive technology use, and recent studies have shown encouraging outcomes. According to their study, children showed gains in many spheres of functioning, including daily life skills and social interactions. Artificial intelligence tools have become increasingly creative within the niche of mobility aids. According to Zdravkova et al.⁸, technology that uses artificial intelligence to support human activities offers special opportunities for attaining more autonomy and involvement in daily life. This is especially important considering Sabet et al.‘s⁹ formulation of mobility technology as a fundamental right for children, which should be made available to all since their developmental paths largely depend on such technology. These technologies are becoming increasingly important instruments for advancing what Santoro et al.¹⁰ define as functional independence, that is, physical mobility and the capacity to go about daily tasks.

This study aimed to evaluate the effect of AI-powered assistive technologies on the mobility and daily living activities of children with Down syndrome. In particular, we investigated how these instruments affect children’s general quality of life, physical mobility, and independence in daily tasks within a Saudi cultural setting.

The study used a quantitative approach using a thorough questionnaire sent to parents and guardians of children with DS using AI-powered assistive tools. This method considers cultural elements, particularly in the Gulf region¹¹, while building on current frameworks for disability inclusion through artificial intelligence¹². Respecting and valuing these cultural elements in the application of AI tools is vital because they significantly affect the acceptance and efficiency of these technologies.

If used with respect for cultural and family settings, artificial intelligence-powered assistive tools can greatly improve the mobility and daily living capacity of Saudi Arabian children with Down syndrome. This study adds to the growing body of research on how technology can help people with Down syndrome^13,14 by examining how well AI tools work to help people move around and with daily tasks.

The paper is structured as follows. It starts with a review of the literature on assistive technologies driven by artificial intelligence adopted for Saudi Arabian children with Down syndrome. Combining a quantitative survey with qualitative interviews, the methodology section presents the mixed-methods approach. Following a review of important challenges, adoption obstacles, and policy consequences, the Results section shows how artificial intelligence impacts mobility, communication, and daily assistance. Recommendations for improving AI access and integration in the Saudi healthcare system will be included in the paper.

Literature review

This section reviews the existing literature on Down syndrome, mobility and daily assistance challenges, AI-powered assistive tools, and the adoption of AI in Saudi Arabia. It also highlights the gaps in the current research to establish the rationale for this study.

Down syndrome: prevalence and societal perspectives

Down syndrome (DS) is a genetic disorder caused by an extra copy of chromosome 21, which presents with developmental and medical difficulties. With variations influenced by socioeconomic and genetic elements, the worldwide prevalence is estimated to be 1 in 1000 to 1100 live births^15,16. Consanguineous marriages, which increase the risk of congenital abnormalities including DS^17,18, make the condition more common in Saudi Arabia.

Down syndrome management in Saudi Arabia is changing under the influence of cultural attitudes and healthcare development. Early intervention initiatives consistent with Saudi Vision 2030 emphasise mother support and rehabilitation services¹⁹. However, often resulting in delayed access to specialised care, cultural views greatly influence diagnosis, intervention timing, and treatment decisions¹¹. Furthermore, influencing families’ inclination to seek timely medical or educational interventions are society’s stigma and limited knowledge, which cause stress among them⁴. This emphasises the need for thorough, culturally appropriate plans to help children with DS and their families obtain the required medical attention and educational tools.

Mobility and daily assistance challenges in down syndrome

Children with Down syndrome face significant mobility and daily living challenges due to motor development delays, muscle hypotonia, and postural control deficits². These limitations affect their independence, participation in daily activities, and overall QoL³. Research has stressed the need for early physical therapy and organised activity programs to reduce these difficulties²⁰.

The effects go beyond the child to carers, who frequently suffer more stress and a lower quality of life because of the continuous support needed¹⁰. Interventions, including community-based rehabilitation and specialised therapeutic programs, have demonstrated how well they improve motor skills and lower carer load^21,22.

New technologies present interesting ways to improve daily assistance and mobility. Smart home technologies, wearable devices, and socially assistive robotics have shown good results in fostering autonomy, involvement, and quality of life in children with disabilities²³. Nevertheless, despite their promise, limitations on cost, accessibility, and cultural adaptation restrict the general application of these ideas.

Children with Down syndrome also face physical and social obstacles that limit their movement. Emphasising the need for mobility equity, Sabet et al.⁹ supported laws that allow access to assistive technologies. Social misunderstandings regarding the capacity of children with DS may cause them to be excluded from mobility-enhancing treatments, thus supporting dependence²⁴. These elements underline how urgently inclusive policies and interventions are catered to particular requirements of children with DS.

AI-powered assistive tools: transforming daily assistance

Artificial intelligence (AI) has profoundly affected the evolution of assistive technologies by offering creative ideas that improve the mobility, communication, and independence of people with disabilities. Using computer vision, natural language processing (NLP), and machine learning, AI-driven systems produce adaptive support tools based on personal needs²⁵. Children with Down syndrome especially benefit from these technologies because they address important challenges with mobility, speech, social interaction, and daily independence.

Many artificial-intelligence-powered assistive devices have shown encouraging results in helping children with Down syndrome. Wearable AI-based mobility aids help with balance and walking coordination, enabling children to negotiate their surroundings⁸. AI-driven speech recognition and NLP tools provide efficient communication support for children with speech difficulties, enabling nonverbal users to convey their ideas through text or synthesised speech^26–28. Through controlled engagement, socially assistive robots offer interactive learning opportunities and therapy sessions, supporting the acquisition of cognitive and motor skills²⁹. AI has also been integrated into smart home systems, where automated features enable daily routines, including turning lights on and off or reminding users to complete self-care tasks, thereby enhancing their autonomy²³.

Despite these promising technological developments, the translation of AI-assistive tools into widespread practical adoption faces significant global challenges. Recent studies from diverse international contexts have revealed similar adoption barriers, suggesting that implementation challenges extend beyond national boundaries. Boot et al.³⁰ conducted a comprehensive systematic review identifying financial constraints, lack of awareness, and inadequate assessment as primary obstacles to global assistive technology access, which are particularly pronounced in low- and middle-income countries. Cross-cultural research has documented comparable patterns across different regions. Toyokawa et al.³¹ identified infrastructure limitations and cultural adaptation challenges in AI-assisted learning systems for children with special needs in Japan, while broader systematic reviews have confirmed that financial and technical barriers persist across diverse healthcare systems³². Recent comprehensive analyses of AI-driven assistive technologies have emphasised that cost, accessibility, and cultural adaptation remain universal challenges, regardless of geographic location^33,34.

Beyond adoption barriers, several technical and ethical problems continue to limit the effectiveness of AI-assisted technologies. Data privacy is one of the main concerns because many AI-driven assistive tools gather and handle private user data. Algorithmic biases also persist, especially when artificial intelligence models are trained on Western-centric datasets that might not reflect linguistic and cultural diversity fairly. This limitation raises questions regarding the equitable accessibility and efficiency of AI-driven assistive tools for children in many regions, including Saudi Arabia³⁵.

Adoption of AI assistive technologies in Saudi Arabia

As part of Vision 2030^1,36, Saudi Arabia is advancing artificial intelligence integration in healthcare, especially in assistive technologies for persons with disabilities. Although artificial intelligence has shown success in daily assistance, patient management, and diagnosis, several obstacles affect its general acceptance³⁷.

Technological infrastructure differs depending on where one lives; urban centres such as Riyadh and Jeddah have advanced systems, while rural areas suffer access restrictions³⁸. Adoption is also influenced by cultural impressions, as some healthcare professionals and carers are still reluctant to make AI-driven decisions³⁹. Furthermore, restricting access, despite increasing government subsidies, is a financial constraint, including the high cost of AI tools and limited insurance coverage⁵.

The lack of localised AI solutions is a major obstacle because many assistive tools are made for Western environments and do not completely fit Arabic language processing or Saudi cultural standards¹¹. Dealing with these issues requires public awareness campaigns, financial incentives, and the creation of culturally fit AI technologies to advance fair adoption and long-term sustainability.

Research gaps

Although assistive technologies driven by artificial intelligence have shown promise worldwide, important research gaps exist regarding their accessibility and efficacy for Saudi Arabian children with Down syndrome. One of the main gaps is the lack of localised research evaluating AI technologies in Saudi socioeconomic and economic settings, which restricts the understanding of their practical relevance.

The lack of culturally adapted AI solutions presents another difficulty because many current technologies are meant for Western healthcare models and might not fit Arabic language needs or regional caregiving practices³. Furthermore, under-studied ethical issues, including data privacy, security risks, and algorithmic bias, affect trust and acceptance⁸.

With higher-income families having better access to assistive technologies than lower-income households, socioeconomic differences also affect AI adoption³⁸. Policymakers, healthcare providers, and artificial intelligence developers depend on filling these gaps to produce more inclusive, efficient, and relevant AI-driven support systems. This study provides insights for future research and policy development, thus helping close these gaps by assessing the effects of AI on mobility and daily assistance within Saudi Arabia’s particular setting.

Materials and methods

We used both quantitative and qualitative research methods to examine how AI-powered assistive tools affect the daily lives and mobility of children with Down syndrome in Riyadh and Al-Jouf, Saudi Arabia. The research design included both a structured survey and in-depth interviews with carers to determine their experiences and challenges with using these technologies. A structured survey collected numerical data on adoption patterns and perceived effectiveness. The study used a variety of data collection methods to obtain a fuller picture of how useful artificial intelligence tools are, determine what factors affect adoption, and compare the experiences of people who use AI tools and those who do not. Previous studies have stressed the need for both quantitative and qualitative data to fully understand the different aspects of how people use technology in their daily lives^40,41. This methodological approach is consistent with these findings.

Participants were categorised into AI users (n = 47) and non-users (n = 76) based on their current use of AI-powered assistive tools for at least three months. Participants were grouped based on the self-reported use of AI tools; thus, the grouping was not randomised. This method made it possible to practically compare those who have embraced artificial intelligence with those who have not.

Study design and ethical considerations

This research was structured as a cross-sectional, comparative study involving both survey data collection and qualitative interviews. The survey was designed to assess mobility, daily task challenges, AI tool effectiveness, and barriers to adoption. The interviews provided in-depth information about the real-life experiences of AI tool users.

The Local Committee of Bioethics (LCBE) at Jouf University provided ethical approval prior to conducting the study (Reference: HAP-13-S-001). Written informed consent was obtained from one parent or legal guardian of all participants under 18 years of age. Additionally, assent was obtained from children aged 12–18 years who demonstrated the cognitive capacity to understand the study purpose and procedures. For children under 12 years of age or those with significant cognitive impairments preventing informed assent, parental consent alone was deemed sufficient, as approved by the Ethics Committee. Participants were informed about the study objectives, data confidentiality measures, and the voluntary nature of their participation through accessible language appropriate to their comprehension level. In adherence to ethical guidelines, participant anonymity was maintained, and the data were securely stored to protect privacy. This study followed the principles of the Declaration of Helsinki to conduct research involving human participants.

Participant recruitment and sampling

This study employed specific inclusion and exclusion criteria to ensure appropriate participant selection and group comparability. For the AI user group, participants were required to demonstrate the current use of at least one AI-powered assistive device for a minimum of three months, ensuring stable adoption rather than temporary trial usage. Children must have been diagnosed with Down syndrome with confirmed karyotype and be currently aged 3–18 years, with their primary carers being Arabic-speaking residents of either the Riyadh or Al-Jouf regions.

The exclusion criteria for the AI user group were designed to eliminate the potential confounding factors. Families with only temporary or trial use of AI devices for less than three months were excluded to ensure committed usage patterns. Those exclusively using non-AI assistive devices, such as conventional wheelchairs or basic communication boards, were also excluded to maintain a clear group distinction. Children with severe medical complications requiring intensive care were excluded, as these conditions could confound the assessment of the effectiveness of technology use. Finally, families that were unable to provide appropriate informed consent were excluded according to the ethical requirements.

The non-AI user group served as a comparison cohort with similar inclusion criteria, requiring children diagnosed with Down syndrome aged 3–18 years with confirmed karyotype, Arabic-speaking primary carers, and residence in the study regions. The critical distinction is the complete absence of current AI-powered assistive technology use. Exclusion criteria included current use of AI-powered assistive devices, severe medical complications requiring intensive care, and inability to provide informed consent.

Data collection and instrumentation

Survey development

A comprehensive Arabic language questionnaire was developed following extensive literature review and consultation with experts in assistive technology, disability studies, and rehabilitation sciences. The instrument development process prioritises cultural appropriateness and linguistic accuracy in the Saudi context.

The questionnaire had rigorous validity and reliability testing to ensure its psychometric soundness. Content validity was assessed through expert reviews conducted by five specialists in disability technology and special education who evaluated the relevance, clarity, and comprehensiveness of all survey items. A pilot study involving 15 carers of children with Down syndrome was subsequently conducted to refine the question clarity and optimise the survey structure based on participant feedback and comprehension patterns.

Internal consistency reliability was evaluated using Cronbach’s alpha coefficients, which demonstrated strong reliability across all the survey domains. The daily challenges subscale achieved a Cronbach’s alpha of 0.85, whereas the AI tool effectiveness subscale demonstrated excellent internal consistency (α = 0.87). The overall questionnaire reliability was confirmed, with α = 0.83, indicating satisfactory internal consistency for the complete instrument.

Interview protocol and procedures

Semi-structured interviews were conducted using a carefully designed guide that balanced standardisation with flexibility for an in-depth exploration of participant experiences. The interview protocol, consisting of six primary questions and accompanying probe questions (detailed in Table 1), was designed to elicit comprehensive responses across key research domains.

Table 1.

Interview items.

Question	Purpose
What motivated you to adopt AI-powered assistive technologies for your child?	Understand the factors driving AI adoption and parental expectations.
What specific AI assistive tools does your child use, and how have they impacted daily life?	Identify the most beneficial AI tools and assess their perceived effectiveness in enhancing mobility, communication, or independence.
What challenges have you faced in using AI-powered assistive technologies?	Explore barriers to effective use, including financial constraints, technical difficulties, usability concerns, and behavioral adaptation challenges.
How has your child responded to AI assistive technologies over time?	Assess the adaptation process, engagement levels, and ease of integrating AI tools into daily routines.
What factors influenced your decision to continue or discontinue using AI-powered assistive tools?	Identify key motivators and deterrents affecting long-term AI tool usage.
What improvements would you suggest for AI assistive technologies?	Gather parental recommendations for enhancing AI usability, accessibility, and affordability.

The interview guide systematically addressed key research domains. Initial questions explored adoption motivations and drivers, seeking to understand the specific challenges carers hoped to address and the influences that shaped their decision-making process. Subsequent questions examined the practical impact of AI tools on daily life, encouraging participants to describe typical usage patterns and identifying both expected and unexpected benefits. Implementation challenges were thoroughly explored with probe questions designed to uncover technical difficulties, coping strategies, and support needs.

The protocol also assessed child adaptation processes over time by examining initial reactions, acceptance patterns, and factors influencing successful technology integration. Sustainability factors were investigated through questions about continued use motivations and potential discontinuation triggers, while the final section gathered enhancement recommendations to inform future technological development and accessibility improvements.

The interview procedures were standardised to ensure consistency across all data collection sessions. Each interview lasted between 45 and 60 min and was conducted at the participants’ preferred location, typically their home or clinic, to maximise comfort and facilitate open dialogue. Digital audio recording was employed with explicit participant consent, and all interviews were conducted in Arabic, with subsequent professional transcription to maintain linguistic accuracy. Seven interviews were conducted, representing users of different AI tool categories and demographic backgrounds. While this sample size was modest, the interviews provided valuable insights into key themes regarding user experiences, implementation challenges, and the perceived benefits of AI-assisted technologies.

Statistical and qualitative data analysis

Quantitative analysis

Quantitative data analysis was conducted using SPSS Version 28.0. Independent t-tests were used to compare AI users and non-users on continuous variables, while chi-square tests were used to examine categorical associations. Levene’s test verified the equality of variance assumptions for parametric testing. The effect sizes were calculated using Cohen’s d, with 0.2, 0.5, and 0.8 representing small, medium, and large effects, respectively. Confidence intervals (95%) were reported for mean differences, and Cronbach’s alpha coefficients were used to assess the scale reliability.

G*Power analysis determined that 100 participants were needed to detect medium effect sizes (d = 0.5), with 80% power at α = 0.05. The final sample of 123 families provided adequate statistical power for group comparison.

Qualitative analysis

The qualitative analysis followed Braun and Clarke’s six-phase thematic analysis. The process began with familiarisation through repeated reading of the interview transcripts and initial note-taking. Systematic coding was conducted using the NVivo 14 software with an inductive approach, allowing themes to emerge from the data. A line-by-line analysis identified meaningful segments that were coded and clustered into potential themes.

Theme review occurred at two levels: against coded extracts to verify coherence and against the entire dataset to ensure representativeness. Clear theme definitions were developed, and compelling extracts were selected to illustrate the findings.

Inter-rater reliability was established through independent coding of 30% of the transcripts by two researchers, achieving Cohen’s kappa = 0.78, indicating substantial agreement. Discrepancies were resolved through discussion and consensus building.

Result

This section presents the results of a study that examined how AI-powered assistive tools affected the mobility and daily tasks of children with Down syndrome in Saudi Arabia. The study used both quantitative survey data (n = 123) and qualitative insights from seven interviews with the caregivers.

Quantitative results

Table 2 provides an overview of the demographic characteristics of the participants, offering insights into the distribution of AI-assistive tool users compared to non-users. In terms of gender distribution, the proportions were relatively balanced, with males representing 60% of AI users and 58% of non-users. This suggests that gender does not play a significant role in the adoption of AI-assistive tools.

Table 2.

Demographic characteristics of participants.

Category	Subcategory	AI users (n = 47)		Non-AI users (n = 76)		Test statistic	p-value
		Number	Percentage	Number	Percentage
Gender	Male	28	60%	44	58%	χ² = 0.062	0.803
	Female	19	40%	32	42%
Age	Less than 5 years	10	21%	19	25%	χ² = 1.247	0.742
	5–10 years	17	36%	28	37%
	11–15 years	15	32%	18	24%
	More than 15 years	5	11%	11	14%
Parent’s education level	Less than High School	3	6%	13	17%	χ² = 15.743	0.003**
	High School	5	11%	17	22%
	Diploma	8	17%	23	30%
	Bachelor’s	18	38%	14	18%
	Postgraduate	13	28%	9	12%
Monthly income	Less than 5000 SAR	3	6%	7	9%	χ² = 22.891	< 0.001**
	5000–10,000 SAR	10	21%	33	43%
	10,001–15,000 SAR	19	40%	10	13%
	More than 15,000 SAR	15	32%	8	11%
Region	Riyadh	31	66%	48	63%	χ² = 0.143	0.705
	Al-Jouf	16	34%	28	37%

Chi-square tests revealed significant differences in parental education and monthly income between groups. p < 0.05, **p < 0.01.

Regarding the age distribution of children with Down syndrome, the majority fall within the 5–10 years category (36% of AI users and 37% of non-users), followed by the 11–15 years category (32% and 24%, respectively). This indicates that AI adoption is widespread across age groups. However, a slightly higher percentage of younger children (under five years) was found among non-users (25%), which suggests that families with younger children are more hesitant to integrate AI tools at an early stage.

Table 2 also reveals notable differences in educational levels between the two groups. AI users tended to have higher educational qualifications, with 38% holding a bachelor’s degree and 28% possessing postgraduate qualifications. In contrast, non-users were more concentrated in lower education levels, particularly at the diploma (30%) and high school (22%) levels. This pattern suggests that higher education may be associated with greater awareness of and willingness to adopt AI-powered assistive tools.

Economic status also appears to influence AI adoption. A significant proportion of AI users fell within the income bracket of 10,001–15,000 SAR (40%), with 32% earning more than 15,000 SAR. In contrast, 43% of non-users belong to the lower-income bracket of 5,000–10,000 SAR. This difference indicates that financial constraints may be a limiting factor in AI adoption because families with higher income levels may have greater access to these technologies.

Table 2 shows that there are only small differences in where AI users and non-users live. Most of the participants live in Riyadh (66% of AI users and 63% of non-users), while the rest are from Al-Jouf. This means that cities like Riyadh may have easier access to AI tools that can help people than less developed areas. This could change the number of places that use these technologies and how widely they are used.

Participant characteristics

This study recruited 123 parents and carers of children diagnosed with Down syndrome, comprising 47 families actively using AI-powered assistive tools and 76 families not using such technologies. While participants were not randomly assigned to groups (as AI usage was self-selected), we ensured group comparability by matching on key child characteristics, including age, cognitive functioning level, physical mobility status, and associated medical conditions. Demographic differences between groups (parental education and income) were treated as covariates in subsequent analyses to control for potential confounding effects.

The baseline characteristics of the participating children were assessed to ensure group comparability (Table 3). Cognitive functioning levels were distributed similarly across both groups, with approximately half of the participants in each group classified as having moderate intellectual disability (IQ 35–49). Mild intellectual disability (IQ 50–70) was observed in 32% of AI users compared to 29% of non-users, while severe intellectual disability (IQ 20–34) affected 17% and 21%, respectively. Chi-square analysis revealed no significant difference in the cognitive functioning distribution between the groups (χ² = 0.625, p = 0.732).

Table 3.

Child characteristics and clinical profiles.

Characteristic	AI users (n = 47)	Non-users (n = 76)	Test statistic	p-value
Cognitive Functioning Level			χ² = 0.625	0.732
Mild intellectual disability (IQ 50–70)	15 (32%)	22 (29%)
Moderate intellectual disability (IQ 35–49)	24 (51%)	38 (50%)
Severe intellectual disability (IQ 20–34)	8 (17%)	16 (21%)
Physical Mobility Status			χ² = 0.937	0.627
Independent walking	28 (60%)	43 (57%)
Assisted walking (with support)	15 (32%)	26 (34%)
Non-ambulatory (wheelchair dependent)	4 (8%)	7 (9%)
Associated Medical Conditions
Congenital heart disease	12 (26%)	19 (25%)	χ² = 0.009	0.925
Hearing impairment	8 (17%)	14 (18%)	χ² = 0.033	0.856
Vision problems	6 (13%)	11 (14%)	χ² = 0.056	0.813
Hypothyroidism	11 (23%)	20 (26%)	χ² = 0.134	0.714
Communication Abilities			χ² = 1.125	0.569
Verbal communication	29 (62%)	44 (58%)
Limited verbal with gestures	14 (30%)	26 (34%)
Nonverbal communication	4 (8%)	6 (8%)

Chi-square tests revealed no significant differences between groups, confirming successful matching for baseline characteristics.

Physical mobility status also demonstrated comparable distributions between the groups. Independent walking was achieved by 60% of the children in the AI user group and 57% in the non-user group. Approximately one-third of the participants in both groups required assisted walking with support (32% AI users, 34% non-users), while non-ambulatory status requiring wheelchair dependence was observed in 8% and 9% of participants, respectively. Statistical analysis confirmed no significant differences in the mobility status (χ² = 0.937, p = 0.627).

The associated medical conditions commonly observed in Down syndrome were similarly distributed between the groups. Congenital heart disease affected approximately one-quarter of the participants in both groups (26% AI users and 25% non-users, p = 0.925). Hearing impairments were present in 17% and 18% of AI users and non-users, respectively (p = 0.856), while vision problems affected 13% and 14% of AI users and non-users, respectively (p = 0.813). Hypothyroidism was diagnosed in 23% of AI users compared to 26% of non-users (p = 0.714). No medical conditions showed statistically significant differences between the groups.

Communication abilities were assessed across three categories, with verbal communication being the most common ability in both groups (62% of AI users, 58% of non-users). Limited verbal communication with gesture support was observed in 30% of AI users and 34% of non-users, whereas nonverbal communication affected 8% of participants in both groups. Chi-square analysis revealed no significant differences in communication ability (χ² = 1.125, p = 0.569).

A 5-point Likert scale was used to compare daily challenges faced by children with Down syndrome between AI users and non-users. The results are presented in Table 4. The results show that people who do not use AI face many more problems in every area. The largest differences were observed when doing household tasks (M = 4.34, SD = 0.85, p = 0.001), talking and communicating (M = 4.21, SD = 1.08, p = 0.002), and moving around (M = 4.13, SD = 1.21, p = 0.004). These statistically significant differences are consistent with the hypothesis that AI-powered assistive technologies may help mitigate challenges in these areas, though causality cannot be established based on cross-sectional data.

Table 4.

Statistical analysis of daily challenges: likert ratings and group comparisons.

Daily challenge	AI users (n = 47)							Non-AI users (n = 76)							T-statistic	P-value	Levene’s statistic	Levene’s P-value	Cohen’s d
	1	2	3	4	5	M	SD	1	2	3	4	5	M	SD
Mobility and movement	5	7	13	15	7	3.25	1.42	3	5	9	21	38	4.13	1.21	− 2.97	0.004**	1.65	0.202	0.651
Social interaction	4	6	12	16	9	3.42	1.39	2	6	10	19	39	4.14	1.17	− 2.80	0.006**	2.10	0.151	0.581
Communication and speech	3	8	14	14	8	3.34	1.28	2	4	11	18	41	4.21	1.08	− 3.21	0.002**	1.98	0.183	0.721
Performing household tasks	2	7	11	17	10	3.55	1.22	1	3	9	20	43	4.34	0.85	− 3.68	0.001**	2.35	0.122	0.832
Learning and educational activities	4	5	15	15	8	3.38	1.30	3	6	13	17	37	4.03	1.32	− 2.15	0.032*	1.89	0.173	0.491
Personal care (eating, hygiene)	3	6	13	13	12	3.53	1.39	2	5	10	19	40	4.18	1.12	− 2.45	0.023*	2.00	0.164	0.543

This table presents the distribution of daily challenges faced by children with Down syndrome, comparing AI users (n = 47) and non-AI users (n = 76) based on a 5-point Likert scale (1 = No Challenge, 5 = Very Significant Challenge). Mean (M) and Standard Deviation (SD) values indicate that non-AI users report greater difficulties across all categories, particularly in mobility, communication, and performing household tasks, compared to AI users. *p < 0.05, **p < 0.01.

Among AI users, the reported difficulty levels were lower, with mean scores ranging from 3.25 to 3.55, indicating moderate challenges rather than severe ones. The lowest reported challenge among AI users was mobility and movement (M = 3.25, SD = 1.42), suggesting that walk-assist devices may have a positive impact on enhancing mobility and coordination. Additionally, AI users reported lower challenges in communication (M = 3.34, SD = 1.28), which are likely to reflect the benefits of AI-powered communication apps in improving speech and interaction.

Statistical analysis supported these observations. The T-test results (p < 0.05) showed that AI users and non-users were significantly different in all categories. Levene’s test results confirmed that the variances between the two groups were comparable, thus ensuring the reliability of the findings. Cohen’s d effect sizes range from 0.49 to 0.83, which indicates moderate to large effects. The most noticeable effect was observed when AI was used to do housework (d = 0.83), which supports AI’s role in encouraging independence.

Despite these benefits, both groups continued to report moderate challenges in personal care (M = 3.53 for AI users, M = 4.18 for non-users) and learning activities (M = 3.38 for AI users, M = 4.03 for non-users). This means that AI-based assistive technologies may help in some ways but may not completely address challenges in self-care and educational activities.

Figure 1 demonstrates a consistent and statistically significant pattern wherein AI technology users report substantially lower daily challenge severity across all measured domains than non-users. The magnitude of the difference was particularly pronounced in household tasks (AI users: M = 3.55, non-AI users: M = 4.34, p < 0.001) and communication challenges (AI users: M = 3.34, non-AI users: M = 4.21, p < 0.01). These findings suggest that AI-assistive technologies may provide meaningful functional benefits in reducing perceived challenge severity across multiple life domains. The consistency of this pattern across diverse functional areas, from mobility and social interaction to personal care, indicates that AI technology adoption may have a broad positive impact on daily functioning in children with Down syndrome.

Fig. 1.

Comparison of Daily Challenge Severity Between AI Users and Non-AI Users. Comparative analysis of daily challenge severity ratings across six functional domains between AI technology users (n = 47) and non-users (n = 76). Challenge severity was measured on a 5-point Likert scale (1 = No Challenge, 5 = Very Significant Challenge). Error bars represent standard deviations. Asterisks indicate statistical significance levels: *p < 0.05, **p < 0.01. AI users consistently reported lower challenge severity across all domains, with statistically significant differences observed in all categories (p < 0.05).

Figure 2 shows AI-powered assistive tool distribution among participants, with smart communication apps most commonly used (36%), followed by AI-powered learning apps (26%) and smart home assistants (17%). These patterns suggest carers prioritise tools enhancing communication, education, and daily task management, aligning with reported challenges. Lower adoption rates for walk-assist devices (9%) and social-assistive robots (6%) may reflect cost barriers, limited accessibility, or effectiveness uncertainty. Despite mobility being a major challenge among non-AI users, low walk-assist device adoption suggests access or awareness limitations. Similarly, social assistive robots remain underutilised despite their interaction benefits, possibly due to high costs or limited availability in the study area. The custom tools category (6%) indicates some families use specialised solutions tailored to individual needs, highlighting assistive technology adoption’s individualised nature, where standard solutions may not meet each child’s unique requirements.

Fig. 2.

Distribution of AI-powered assistive technology Categories Among Study Participants. Frequency distribution of AI assistive tools used by families with Down syndrome children (n = 47). Bars show number of users and percentage for each technology category, ranked by adoption rate. Smart communication apps were most commonly used (17 users, 36%), while social assistive robots and custom solutions had the lowest adoption (3 users each, 6%). Representative examples of tools within each category are shown in parentheses.

Table 5 demonstrates patterns in AI-powered assistive technology adoption sustainability, with usage duration following a declining trajectory, where only 13% of participants maintained usage beyond 24 months compared to 38% in the 6–12 month range, indicating significant long-term retention challenges. Implementation occurs predominantly in home environments (89%), with tool-specific variations reflecting functional characteristics. Walk-assist devices show the highest clinical involvement (25%) because of their medical nature, while smart home assistants and social assistive robots demonstrate exclusive domestic usage (100%), and communication apps remain largely home-centred (94%). The statistically significant relationship between usage duration and perceived effectiveness (F(3,43) = 4.567, p = 0.008, η² = 0.242) suggests that sustained engagement enhances outcomes, with an effectiveness variance attributable to usage duration of approximately 24%.

Table 5.

AI tool usage duration and implementation settings.

AI tool category	Usage duration				Primary usage setting
	3–6 months	6–12 months	12–24 months	> 24 months	Home	Clinic/Hospital	Mixed
Walk-assist devices (n = 4)	1 (25%)	2 (50%)	1 (25%)	0 (0%)	3 (75%)	1 (25%)	0 (0%)
Smart communication apps (n = 17)	4 (24%)	7 (41%)	4 (24%)	2 (11%)	16 (94%)	0 (0%)	1 (6%)
Smart home assistants (n = 8)	2 (25%)	3 (38%)	2 (25%)	1 (12%)	8 (100%)	0 (0%)	0 (0%)
AI-powered learning apps (n = 12)	3 (25%)	4 (33%)	3 (25%)	2 (17%)	11 (92%)	0 (0%)	1 (8%)
Social assistive robots (n = 3)	1 (33%)	1 (33%)	1 (33%)	0 (0%)	3 (100%)	0 (0%)	0 (0%)
Other tools (n = 3)	1 (33%)	1 (33%)	0 (0%)	1 (33%)	2 (67%)	0 (0%)	1 (33%)
Total (n = 47)	12 (26%)	18 (38%)	11 (23%)	6 (13%)	42 (89%)	1 (2%)	4 (9%)

Usage duration is measured from initial adoption to study participation. A one-way ANOVA revealed significant differences in perceived effectiveness based on usage duration (F(3,43) = 4.567, p = 0.008, η² = 0.242). Settings were categorized as home (primary residence), clinic/hospital (healthcare facilities), or mixed (both settings regularly).

Table 6 presents the perceived effectiveness of AI-assistive tools across four key areas: mobility, daily independence, social interaction, and overall quality of life. The ratings, based on a 5-point Likert scale, indicate moderate to high effectiveness, with mean scores ranging between 3.68 and 3.85.

Table 6.

Perceived effectiveness of AI assistive tools.

Effectiveness category	AI users (n = 47)							One-sample t-test vs. neutral (3.0)
	1	2	3	4	5	Mean	SD	t-statistic	p-value
Increasing daily independence	1	3	11	19	13	3.85	0.93	6.24	< 0.001**
Improving overall quality of life	2	4	9	18	14	3.80	1.17	4.69	< 0.001**
Improving mobility	2	4	10	17	14	3.78	1.18	4.53	< 0.001**
Enhancing social interaction	3	5	10	15	14	3.68	1.40	3.33	0.002**

Perceived effectiveness ratings on a 5-point Likert scale (1 = Not Effective, 5 = Very Effective). One-way repeated-measures ANOVA revealed significant differences between the effectiveness domains (F(3,138) = 2.847, p = 0.040, η² = 0.058). Post-hoc tests indicated that daily independence was rated significantly higher than social interaction (p = 0.032). *p < 0.01.

The category with the most votes (M = 3.85, SD = 0.93) was “increasing daily independence.” This means that AI tools, especially smart home assistants and AI-powered learning apps, are seen as helpful in helping children perform daily tasks with less help from carers. The relatively low standard deviation (0.93) indicated greater agreement among participants regarding the effectiveness of these tools.

Improving mobility (M = 3.78, SD = 1.18) also received a relatively high score, reflecting the positive impact of walk-assist devices for children with Down syndrome. A higher standard deviation (1.18) suggests that the responses varied more. This could be because people have different mobility needs or difficulties accessing these devices.

Enhancing social interaction (M = 3.68, SD = 1.40) had the lowest mean score, indicating that, while smart communication apps and social assistive robots provide some benefits, their effectiveness may vary depending on the child’s individual needs and the tool’s usability. The higher standard deviation (1.40), which also suggests that user experiences are different, could be due to factors such as how responsive the child is to AI-driven social interactions or how difficult it is for them to use these tools effectively.

Improving the overall quality of life (M = 3.80, SD = 1.17) received a slightly higher rating, reflecting the general positive impact of AI-assistive tools across multiple domains. Moderate standard deviation indicates some variability in perceived benefits, which may be influenced by the extent of tool adoption and integration into daily routines.

Table 7 highlights the key challenges in using AI assistive tools, with high cost (74.46%) being the most significant barrier, limiting accessibility for many families. Lack of technical support (38.29%) and difficulty adapting to technology (31.91%) indicated that usability and training remain concerns. Additionally, software issues (25.53%) and device incompatibility (21.27%) suggested that not all AI tools meet individual needs. These findings emphasise the need for financial assistance, better user support, and improved customisation to enhance the effectiveness and accessibility of AI-assistive tools.

Table 7.

Challenges in using AI assistive tools.

Challenge	Users (n = 47)	Percentage (%)	95% CI	Binomial Test vs. 25%*	p-value
High cost	35	74.46%	[59.7%, 86.1%]	23.10	< 0.001**
Lack of technical support	18	31.91%	[24.5%, 53.6%]	2.06	0.041*
Difficulty adapting to technology	15	38.29%	[19.1%, 47.1%]	1.07	0.287
Software or usability issues	12	25.53%	[13.9%, 40.4%]	0.09	0.931
Device incompatibility with child’s needs	10	21.27%	[10.7%, 35.8%]	-0.61	0.543
Other (custom responses)	7	14.89%	[6.2%, 28.3%]	-1.71	0.089

Expected baseline of 25% based on literature on general technology adoption challenges. *p < 0.05, *p < 0.01.

Figure 3 presents a comparative analysis of adoption barriers and corresponding solution preferences among families currently not using AI-assisted technologies. Visualisation reveals important insights into intervention opportunities by juxtaposing barrier prevalence with solution appeal across the five key domains.

Fig. 3.

Barrier prevalence and solution appeal analysis for AI-powered assistive technology adoption among non-users (n = 76).

Infrastructure and access limitations demonstrated the highest intervention potential, with a net impact on + 16 families (+ 21.1%). While only 22 families (28.9%) cited this as a barrier, 38 families (50.0%) expressed an interest in education and skill development programs, suggesting that targeted training initiatives could address broader adoption challenges beyond immediate infrastructure concerns.

Financial constraints, though affecting the largest proportion of families (38 families, 50.0%), show more modest intervention potential (+ 7 families, + 9.2%) due to the correspondingly high interest in financial support programs (45 families, 59.2%). This indicates that, while financial barriers are widespread, the proposed solutions align well with the identified needs.

Technology dependency fears presented a concerning pattern, being the only category with a negative net impact (-5 families, -6.5%). The limited appeal of custom incentives (10 families, 13.2%) relative to the prevalence of technology fears (15 families, 19.7%) suggests that this barrier may require more innovative intervention approaches that are not captured in conventional solution categories.

Awareness and trust issues and scepticism concerns both demonstrated positive but moderate intervention potential (+ 7 and + 10 families, respectively), indicating that trial-based programs and community validation strategies could effectively address these barriers.

Qualitative results

To gain more profound insights into the experiences of parents using AI-powered assistive tools for their children with Down syndrome, semi-structured interviews were conducted with seven carers (see Table 8). The interviews aimed to explore the perceived benefits, challenges, and factors influencing the adoption and continued use of AI-based assistive technologies. The data were analysed using thematic analysis, which allowed the identification of key themes emerging from the participant responses. The qualitative findings provide additional perspectives that were not fully captured in the quantitative survey, particularly regarding the emotional and psychological impacts of AI tools on carers and their children.

Table 8.

Characteristics of interviewees (N = 7).

Participant ID	Child’s age (years)	AI tool used	Region	Parent’s education level	Monthly income
P1	6	Smart communication app	Riyadh	Bachelor’s	10,000–15,000
P2	10	Walk-assist device	Riyadh	Bachelor’s	15,000+
P3	8	AI-powered learning app	Al-Jouf	Master’s	15,000+
P4	12	Smart home assistant	Riyadh	High School	5,000–10,000
P5	9	Social assistive robot	Al-Jouf	Bachelor’s	10,000–15,000
P6	11	Smart communication app	Riyadh	Diploma	5,000–10,000
P7	7	AI-powered learning app	Al-Jouf	Master’s	15,000+

Perceived benefits of AI assistive tools

Participants generally expressed satisfaction with the impact of the AI-powered assistive tools on their children’s daily activities, communication, and independence. Many parents noted that these technologies provided tangible improvements in mobility, learning, and social interaction, ultimately enhancing their children’s ability to navigate their daily lives more effectively. The most frequently reported benefit was enhanced mobility, particularly among parents whose children use walk-assist devices. These tools not only helped children improve their physical movement but also contributed to greater confidence and willingness to explore their surroundings. One participant highlighted the transformation.

“Since using the robotic walking aid, my child has become more eager to move independently, which was not the case before.” (P2).

In addition to mobility improvements, smart communication apps have been widely valued, especially by parents of nonverbal children. These technologies allow children to express their needs, emotions, and preferences more clearly, reducing frustration for both the child and the carer. Improved communication has a direct impact on family interactions, fostering meaningful engagement between the child and their surroundings. A carer shared:

“Before using the communication app, my child struggled to communicate his needs. Now, he points at the screen, and we understand him much faster.” (P1).

Beyond the direct benefits for children, the qualitative findings also revealed an important new insight that was not captured in the quantitative data: the emotional relief experienced by carers. Several parents have reported that AI tools contribute to a noticeable reduction in caregiving stress. As these technologies enable their children to engage in self-care activities and structured learning independently, parents experience less pressure and mental exhaustion associated with constant supervision. This shift allowed carers to allocate more time to other responsibilities while feeling reassured of their children’s engagement. One mother described this change as follows:

“With the AI learning app, my child stays engaged and enjoys the lessons, giving me time to focus on other tasks.” (P3).

This suggests that the impact of AI tools extends beyond the child, influencing the broader family dynamics and the overall well-being of carers. By reducing the daily burden of caregiving, these technologies create a more balanced environment for both parents and children, demonstrating the secondary but equally important benefit of AI-powered assistive tools.

Challenges in using AI assistive tools

Despite the numerous benefits reported by parents, the adoption and integration of AI-powered assistive tools into daily routines present several challenges. One of the most frequently cited barriers is cost, which limits access to AI tools for many families. The financial burden associated with purchasing and maintaining these technologies was described as a significant obstacle, preventing some families from adopting them or restricting the range of tools that they could afford. One participant explained,

“We had to make financial sacrifices to afford these tools, and not all families can do that.” (P5).

Beyond affordability, technical difficulties have emerged as major concerns. Many parents struggle with frequent software updates, troubleshooting errors, and a lack of adequate customer support, making it difficult to use AI tools effectively. Some reported frustration with the absence of clear guidance or assistance, which made it challenging to resolve technical issues independently. One participant shared their experiences with a smart home assistant, stating:

“When the smart home assistant stopped responding, I had no idea how to fix it, and there was no clear guidance available.” (P4).

This indicates that while AI assistive tools have the potential to significantly improve children’s independence and quality of life, their effectiveness is often contingent on the level of technical support available to carers.

Another major challenge was child adaptation and engagement, which varied significantly across the families. While some children quickly embraced AI tools, others showed resistance or required extended adjustment periods before they could comfortably use the technology. Parents noted that some children, particularly those with sensory sensitivities or aversion to change, initially rejected AI devices and required gradual exposure and encouragement. One participant described their child’s reluctance to use a mobility-assistive device:

“At first, my child refused to wear the mobility device, and it took weeks of encouragement before he started using it comfortably.” (P2).

This finding suggests that the successful integration of AI tools may require structured behavioural strategies, such as gradual exposure, positive reinforcement, and tailored support from therapists or educators. Given that each child responds differently to technology, a one-size-fits-all approach may not be effective, thus emphasizing the need for customizable solutions and adaptive training programs to facilitate a smoother transition.

Overall, the challenges identified by parents highlighted three key areas for improvement: financial accessibility, enhanced technical support, and structured adaptation strategies. Addressing these barriers could significantly improve the long-term effectiveness and usability of AI-powered assistive tools, making them more accessible and beneficial to children with Down syndrome and their families.

Factors influencing adoption and continued use

Several key factors were identified as critical motivators for the adoption and continued use of AI-powered assistive tools. Government financial support was highlighted as a primary enabler, with one parent stating:

“If financial aid were available, I believe more families would adopt these technologies.” (P6).

Additionally, training programs for parents and carers are essential for effective use. Participants emphasised that many AI tools require technical knowledge that parents may lack, making educational support crucial.

Peer recommendations also play a significant role in shaping the adoption decisions. Many carers reported that they initially hesitated to use AI tools but were persuaded by success stories from other families or professionals. One participant noted:

“I was sceptical at first, but after hearing another parent’s experience, I decided to give it a try, and it has been life-changing.” (P7).

A new finding, not evident in the quantitative analysis, was that some parents experienced a shift in perception regarding technology dependence. Initially, there was a concern that AI tools might create over-reliance, but over time, many parents viewed them as complementary rather than replacing parental interaction. One participant expressed the following:

“At first, I worried that my child would become too dependent on the AI assistant, but now I see it as a helpful tool rather than a replacement for family engagement.” (P4).

This suggests that initial hesitation about AI reliance may diminish over time, as carers experience tangible benefits.

These qualitative findings reinforce the survey results, confirming that AI-powered assistive tools significantly enhance mobility, communication, and independence in children with Down syndrome. However, the interviews provided additional insights that were not captured in quantitative data, particularly in relation to the emotional relief experienced by carers and the evolving perceptions of AI reliance.

Although financial constraints and technical barriers remain substantial obstacles, the qualitative results highlight the importance of structured training programs, peer networks, and government financial support in ensuring successful adoption. Addressing these factors could broaden accessibility and improve the long-term impact of AI-powered assistive technologies on both children and their families.

Discussion

This study sought to investigate how AI-driven assistive technologies might affect daily assistance, mobility, and communication in children with Down syndrome in Saudi Arabia. Particularly for young people using walk-assist devices and smart communication apps, the findings indicate that AI tools may contribute to improvements in independence and daily functionality, with varying effects based on the specific tools used and individual characteristics. Consistent with the results of previous studies, these tools appear to assist with mobility coordination, contribute to self-care routines, and may facilitate participation in social events, although individual responses to the technology vary considerably^25,29. However, the study also highlighted ongoing obstacles to adoption, including technical difficulties, financial restrictions, and differences in accessibility, implying that socioeconomic and cultural elements shape how artificial intelligence is included in assisted care.

The current findings confirm and significantly extend previous research in several ways. Our observed effect sizes for mobility improvement (Cohen’s d = 0.65) were larger than those reported in some Western studies, where Zdravkova et al.⁸ documented moderate effects (d = 0.42) for AI-powered mobility aids. However, direct comparisons should be interpreted cautiously due to differences in study designs, populations, and measurement approaches. The observed differences could potentially reflect cultural variations in family engagement patterns, measurement instruments, or selection effects, rather than indicating superior AI effectiveness in Gulf contexts.

Similarly, our communication improvement effect size (d = 0.72) surpasses findings from European contexts (d = 0.51; Krasniqi et al.²⁷, suggesting that AI communication tools may be particularly beneficial for Arabic-speaking children who face additional language-processing challenges with predominantly English-based AI systems. Importantly, our study addressed the critical gap identified in the literature review regarding culturally adapted AI solutions. While Sahoo and Choudhury²⁶ emphasised the potential of AI-driven communication tools, our findings reveal that effectiveness varies significantly based on cultural and linguistic adaptation. The moderate effectiveness ratings for social interaction tools (M = 3.68) compared to mobility aids (M = 3.78) align with Panceri et al.‘s²⁹ observations about variable outcomes in socially assistive robotics; however, our qualitative data provide new insights into cultural factors influencing these differences. The financial barriers documented in our study (74.46% citing high costs) closely mirror the global patterns identified by Boot et al.³⁰, who found cost to be the primary barrier across diverse international contexts. However, our finding that 59.2% of non-users would adopt AI tools with government financial support suggests a higher potential demand than previously documented, possibly reflecting Saudi Arabia’s growing technological infrastructure as part of Vision 2030 initiatives⁵.

The results of this research match earlier ones that stressed the need for artificial intelligence to increase accessibility and independence for people with disabilities^12,25. Western studies have shown that AI-powered communication apps make it much easier for nonverbal children to talk and interact with others. This makes them happier and improves how families talk to each other^26,27. The current study validates these findings, since Saudi carers claimed better awareness of their children’s needs following the application of smart communication technologies. Furthermore, AI-driven mobility aids are acknowledged for their efficiency in improving mobility coordination and autonomy, which is in line with the results of Zdravkova et al.⁸.

However, the results revealed significant differences in AI perception between Saudi Arabia and other regions, particularly regarding accessibility, cultural beliefs, and financial concerns. Unlike studies conducted in high-income countries where government support and insurance coverage facilitate AI adoption, this study found that financial constraints are a major obstacle for Saudi families. Similar to the results of Asem et al.⁵, cost was found to be the main obstacle to AI adoption, particularly because Saudi Arabia has limited insurance coverage for AI-powered devices. It was also discovered that cultural views on AI-driven care and decision-making affected adoption patterns. This is similar to what Crisera et al.³⁹ found about people who do not like AI-based decision-making in healthcare settings.

The significance of culturally modified artificial intelligence solutions for Saudi Arabian children with Down syndrome is another important finding. Many artificial intelligence tools, which are based on Western models and might not fit Arabic language needs or Gulf-region caring practices¹¹, have been designed. The qualitative findings reveal that carers struggle with usability issues and language barriers, thereby supporting earlier studies advocating for the development of localised artificial intelligence³. Furthermore, the lack of culturally inclusive AI models may limit the effectiveness of these technologies in Saudi children.

The study produced surprising results, including the psychological relief felt by carers using AI-assisted technologies. Although most studies on the acceptance of artificial intelligence concentrate on functional advantages for children, this study revealed that carers mentioned lower stress and more freedom in their daily lives when their children interacted with AI-powered learning and communication tools. This supports the more general case that assistive technologies not only increase child autonomy but also reduce carer burden¹⁵.

Variability in child adaptation to artificial intelligence tools is another intriguing result. While some children interacted with AI-powered assistive technologies immediately, others needed more time to adjust or showed initial resistance. This implies that the adoption of artificial intelligence among children with disabilities is not homogeneous, and that effective implementation could call for organised behavioural interventions²². The results of the study showed that overcoming initial resistance to artificial intelligence use depends on gradual familiarisation, tailored instruction, and continuous carer support.

The results highlight the need for targeted legislative action to improve AI availability for Saudi Arabian children with disabilities. Increasing government subsidies and insurance coverage for AI-powered assistive devices could help close the financial gap and promote more general adoption³⁷. Creating training courses for educators and carers would also raise AI literacy and usability, enabling families to include these tools in their daily lives¹⁴.

AI developers should prioritise localisation efforts, given the cultural and linguistic difficulties found in the research, ensuring that assistive technologies are tailored to Arabic-language users and Saudi care settings³. More inclusive artificial intelligence-driven support systems depend on cooperation among technology companies, healthcare providers, and legislators.

Our study extends the current theoretical framework in several meaningful ways. While previous studies have focused primarily on functional outcomes for children^7,12, our discovery of significant carer stress reduction represents a novel secondary benefit not extensively captured in the existing literature. This finding suggests that AI-powered assistive technology adoption should be conceptualised using family systems theory rather than individual-focused models, as the technology’s impact ripples through family dynamics.

The variability in children’s adaptation to AI tools documented in our qualitative findings challenges the assumption of uniform technology acceptance prevalent in the literature on assistive technology. While Almoghyrah²² noted individual differences in educational technology adoption, our data reveal that successful AI integration may require structured behavioural interventions, a finding that bridges assistive technology and applied behaviour analysis literature in novel ways.

Furthermore, our identification of cultural adaptation barriers provides empirical support for AlQahtani and Efstratopoulou’s¹¹ theoretical arguments on Gulf-specific cultural factors in disability interventions. The gap between AI tool availability and cultural appropriateness documented in our study validates calls for localised AI development while providing a specific direction for culturally responsive design principles.

Limitations and future research directions

This cross-sectional study examined AI-powered assistive technology use among 123 families with children with Down syndrome from two Saudi regions (Riyadh and Al-Jouf). The study looked at how families began using AI-powered assistive technology on their own and found that families who used AI tended to have higher education levels (66% had bachelor’s degrees compared to 30%) and higher incomes (72% earned more than 10,000 SAR compared to 24%), which are typical characteristics of early technology users that were considered in the analysis.

Data collection relied on carer-reported outcomes, which capture important real-world functional perspectives, complemented by seven qualitative interviews that offered preliminary conclusions about user experiences. The study assessed diverse AI technologies with varying sophistication and cultural adaptations, reflecting current market availability. The cross-sectional methodology captured associations at a single time point, documenting usage patterns and perceived outcomes rather than causal intervention effects.

Future research should look at how AI is used over a longer period, include more in-depth studies to represent different user groups, and cover various regions in Saudi Arabia with different cultures. Randomised controlled trials would help show how AI use directly affects results, and using objective measures would support the personal feedback we obtain. Research priorities include creating and testing AI tools that are suitable for Arabic-speaking people, studying how well these tools work for different income groups to help make access fair, and doing cost studies to help decide how to use resources effectively. These studies would use the existing findings to improve the use of AI assistive technology in Saudi Arabia and similar cultures.

Conclusion and implications

This study, which used both qualitative and quantitative methods, discovered that using AI-powered assistive technology is linked to improved abilities in mobility, communication, and daily living for children with Down syndrome in Saudi Arabia, showing moderate to large effects. However, the cross-sectional design prevents establishing causality, and observed differences may reflect selection effects or other confounding factors. The research revealed both direct benefits for children and secondary advantages for carers, including reduced stress and increased autonomy in daily responsibilities. However, substantial adoption barriers persist, with financial constraints affecting three-quarters of the potential users.

These findings indicate that successful AI integration requires coordinated policy interventions to address systemic barriers. Government subsidies covering 60–80% of AI-powered assistive technology costs represent the most urgent need, followed by mandated insurance coverage for healthcare-prescribed devices. Regional AI support centres providing technical assistance and carer training would address the substantial gaps in user support currently limiting its effective implementation.

Cultural and linguistic adaptations are equally important to financial accessibility. The limited effectiveness of Western-designed AI tools in Arabic-speaking contexts demonstrates the need for localised development approaches. Technology developers should prioritise Arabic language processing capabilities and culturally appropriate interaction designs to maximise tool effectiveness in Gulf populations.

Future research should focus on three priority areas: longitudinal studies tracking AI technology integration patterns over multiple years, randomised controlled trials comparing specific AI interventions with conventional therapies, and cost-effectiveness analyses for healthcare system implementations. Additional investigations into culturally adapted AI solutions and their comparative effectiveness would inform evidence-based policy development.

The evidence shows that we should be cautiously hopeful about AI’s potential, but we need to actively work on removing barriers to access through coordinated efforts in policy, healthcare, and technology. Without focused actions to lower costs and make AI more culturally relevant, these technologies might worsen existing inequalities instead of helping children with Down syndrome and their families. Without deliberate efforts to reduce financial barriers and improve cultural adaptation, AI-assistive technologies risk exacerbating existing inequities rather than promoting inclusion for children with Down syndrome and their families.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

R.A. and H.B. contributed to the study conception and design. Material preparation, data collection, and analysis were performed by R.A., A.S.A., and S.A. The first draft of the manuscript was written by R.A., while H.B. oversaw the research methodology and statistical analysis. A.S.A. and S.A. contributed to the interpretation of results and provided expertise on educational aspects for children with Down syndrome. All authors reviewed and commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

The authors extend their appreciation to the King Salman Center for Disability Research for funding this work through the Research Group No. KSRG-2024-181.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

The study received Ethical approval was obtained from the Local Committee of Bioethics (LCBE) at Jouf University (Reference: HAP-13-S-001) prior to data collection.