Abstract

Artificial Intelligence (AI) continue to develop at an extraordinary pace across all sectors of industry and society. Employees, Students, Academics, Professionals and Journalists now routinely use these tools as part of their day-to-day activities, sometimes unknown to them. However, there is a large group of individuals who do not have access or limited access to these AI tools due to unemployment, poverty, or disability.

This study will investigate the impact of Privacy and Ethical Considerations that delay the adoption and implementation of these tools despite the clear and obvious benefits for communication, education and employment opportunities for people living with Intellectual Developmental Disabilities (IDD).

This research paper will explore existing literature on the topic while also conducting quantitative research involving relatives or families of people living with IDD’s such as Down Syndrome, ADHD and/or Autism. Through a questionnaire the study intends to firstly identify if families are using AI or ML tools and if not why. Secondly the insights gathered will help define the future topics of research.

Contents

Introduction. 3

Literature Review.. 5

2.1 Introduction. 5

2.2 Privacy Calculus Theory. 5

2.3 Quality of Life. 6

2.4 Research Gap. 8

Methodology. 9

3.1 Study Overview.. 9

3.2 Survey Design. 9

3.3 Sampling Techniques. 9

3.4 Data Collection. 9

3.5 Data Analysis. 10

3.6 Ethical Considerations. 10

Data Analysis & Results. 11

4.1 Data Overview.. 11

4.2 Tests of Normality. 13

4.3 Reliability Testing. 15

Limitations. 15

Future Suggested Research. 16

Conclusion. 17

Acknowledgements. 17

Appendix. 18

References. 34

Introduction

Artificial intelligence (AI) has developed at a phenomenal pace over the last decade, to the extent that people do not realise how integral AI is in their daily life. Computational Social Science is the field dedicated to the use of statistical and computational analysis of different aspects of society and how different phenomena can be analysed and measured for the betterment of society.

However, there are conflicting views on whether this rapid and continuous involvement has the publics best interests at core, such as the risk to vulnerable groups of society due to the increased use of sophisticated algorithms and computational models in AI is leading to heightened concerns regarding privacy according to Jones et al (2024).

This research aims to explore privacy concerns and evaluate whether they impact on the use of Artificial Intelligence tools in various educational and life skills settings for Individuals living with Intellectual Developmental Disabilities (IDD).

This introduction will provide an overview of the Research Project in several ways by first discussing the background and context of where AI tools can be used to improve Quality of Life (QoL) metrics, followed a brief discussion of the research problem, aims, objectives and questions that will be addressed throughout the paper.  The significance of the topic will be identified from the survey results and finally the limitations from the survey will be outlined.

The most common form of IDD is Down Syndrome with a Birth rate of 1 in 700 according to the CDC, while the WHO has analysed data confirming the rate of incidence of Autism is 1 in 100 children. It is quite clear from these statistics that the number of individuals who could benefit from AI in a therapeutic setting is quite high.

Historically people living with IDD were not given the same opportunities and access as others, this is slowly evolving but there is more to be that can be done. People living with an IDD are not always able to communicate their needs, wants, desires, physical condition, emotional wellbeing, and current health status effectively. Each of these areas offers opportunities for AI applications to be developed addressing these needs.

The most common Apps currently available are the See and Learn range which is a combination or Speech and Language Therapy with Occupational therapies from an educational perspective.  Down Syndrome Ireland’s “Good health app” has been developed as a means for people living with Down Syndrome to track their dietary and nutrition habits to share with healthcare professionals and family members.

Numerous studies have investigated Privacy, Bias and Ethical concerns around the use of AI. According to Hutchinson et al (2020), AI has preconceived bias because of the data models it was trained on in two examples of conversation classification and sentiment analysis the models were shown to be more hostile, toxic and negative when disability was mentioned. Almufareh et al (2023), discuss how concerns around the storage of data and the privacy around its collection and retention is a major issue for companies to complying with best practice risk, cyber security protocols while ensuring that they comply with various regulations such as GDPR.

While researchers agree that there are privacy issues around Information Technology covered through extensive research usage through many studies there seems to be limited research on the Privacy and Ethical considerations when it comes to AI and specifically in vulnerable groups of society.

As a result, this Research Paper intends to engage with individuals who have a relative living with IDD to ascertain the level of privacy concerns out there and if this impacts on the usage of AI tools and applications to further the discussion on the current research which is inadequate and unrepresentative of society, specifically targeting the following research question: How do peoples privacy concerns impact on their usage of Artificial Intelligence to achieve better quality of life outcomes for individuals living with Intellectual & Developmental Disabilities?

To achieve this an online survey has been created and shared with various disability groups to capture respondents’ views which is hoped can be used to further contribute to the development of studies around privacy and AI.

It is envisaged that this study can be used to contribute to the development of a body of research looking at how more vulnerable groups of society interact with ICT and AI and that their viewpoints including very specific needs and concerns are factored into future discussions.

The following sections discuss current literature and lack there of in this area to show why this topic is of importance, the research methodology undertaken and then the presentation and analysis of the results. No research paper would be complete without highlighting the limitations within the study and these are documented.

Literature Review

2.1 Introduction

This section will go into detail on current Literature in the area around privacy concerns and the use of AI. The purpose of this section is to discuss current research in Privacy Calculus Theory and identify gaps to support our research question.

Artificial Intelligence (AI) in the last ten years has become a powerful tool in all sectors of industry and society. For people living with Intellectual and Developmental Disabilities (IDD) the potential to enhance Quality of Life (QoL) outcomes is limitless, particularly in areas of Learning, Communication and Life Skills.

However it is clear from current research that Privacy Calculus Theory shows how perceptions are key influencers in decision making around privacy concerns as discussed by Wang et al (2016) and “The Privacy Paradox” where cognitive dissonance also plays a part in influencing individuals decisions as discussed by Whelan et al (2024).

2.2 Privacy Calculus Theory

Privacy Calculus Theory is a Theoretical Framework that focus’ on decision making regarding online data disclosure. Cullnan & Armstrong (1999) described the Privacy Calculus as peoples thought process as they evaluate the trade-off between sharing private information and the perceived risks and benefits associated with the decision.

The theory provides a very valuable medium to understand how individuals make complex decisions in the new AI driven digital world as discussed by Krasnova et al (2012) who identified that trusting beliefs and uncertainty avoidance drive these decisions.

Dinev & Hart (2006) discussed how once the perceived benefits outweighed the perceived risks the risk was deemed acceptable. While Kim et al (2019) went further and demonstrated the above hypothesis in an IoT device study.

Interestingly Haug et al (2020) highlighted that while privacy concerns are at the forefront of peoples mind it does not necessarily relate to a corresponding negative impact on technology adoption.

Princi & Kramer (2020) addressed the issue of whether individuals were comfortable with the control over personal data collection by IoT devices which confirmed that Privacy Calculus could be applied in this area.

Figure 1 Extension of the privacy calculus model to predict intention to use IOT devices (adapted based on Kim & Kim, 2020)

2.3 Quality of Life

Quality of Life (QoL) is multifaceted when it comes to the potential benefits for the IDD communities below studies and projects show from an educational, healthcare and Life Skills perspective that the potential is enormous such as enhanced independence, improved access to education, communication and a better ability to integrate successfully in society.

In September 2022, the National Institute for Health and Care Research in the UK agreed with my assumption and provided funding for the DECODE project which will use Machine Learning to track, analyse and visualise data gathered on the health conditions and comorbidities people with intellectual disabilities live with. The core focus is to build a more joined up approach to provision of health care through proactive planning based on historical data to ensure better health outcomes.

Gupta et al (2022), discussed how Machine Learning can be used as a tool to assist healthcare professionals aided through work done by researchers who have analysed and interrogated data gathered as part of complex medical research projects providing a basis for better identification of potential health related issues by monitoring prevalence of certain conditions such as thyroid and respiratory health in people with intellectual disabilities.

Chao et al (2022), built a Machine learning model to investigate the potential to diagnose Autism in people with intellectual disabilities. Children with Down Syndrome have a 35% chance of a dual diagnosis of Autism. A dual diagnosis like this makes development much more difficult for children and early identification is essential as mentioned above to ensure these kids meet their potential. Through numerous observational techniques Chao and team were able to predict a dual diagnosis with 95% accuracy. Johnson et al (2019) went further and built an AI driven model with tracked and monitored health markers to reduce hospital visits and overall health for people living with IDD.

Engagement and attention are the two areas that educators try to activate in young children to promote a learning environment. Unfortunately for children with intellectual disabilities attention spans are shorter and “engagement, boredom and frustration” comes much quicker. Standen et al (2022), discovered in an albeit small group of children containing 67 subjects that harnessing sensory engagement through AIed tools led to better understanding and outcomes for children. Multi-modal Apps embracing multiple sensory objectives can be built to encourage positive learning environments. Heins et al (2020) built a personalized learning environment for children with autism that showed significant improvements in engagement and attention.

Aggarwal (2018) built a “Support Vector Machine” to identify the differences between typical children and children with intellectual disabilities and had a 97% success rate in identifying the level of intellectual disability in a child from mild to moderate. This distinction in mainstream psychology is notoriously difficult to make and many children are borderline ending up with the wrong diagnosis which will impact future care and schooling options.

A sense of belonging in society is all people want to achieve and this is very difficult when you have an intellectual disability. Unfortunately stereotypes and bias are parts of daily life and impact on opportunities for social inclusion. These bias can be then brought across to Machine Learning depending on the data used for training the applicable models, this is one caution provided by Broda et al (2021) who built a Predictive Machine Learning algorithm which can be used by agencies and policy makers to show the positive impact on the quality of life of individuals with Intellectual disabilities through the tracking of employment and participation in service provision.

Gaurav et al (2022), identified a need to help build a predictive model that measured Quality of Life (QoL) for people living with Intellectual Disabilities as they age. Well-being, Social Inclusion and Independence are unfortunately areas outside the control of Intellectually disabled and as they age and family dynamics change, sheltered housing and residential care becomes inevitable. By using the ML model to monitor the QoL, entities can adjust to ensure better outcomes for people in later life.

It is clear from the above small sample of research papers published that the possibilities and areas that can be positively impacted by AI, Machine Learning and Algorithms for people living with Intellectual Disabilities is limitless and more importantly there is the potential to have an extraordinary impact on lives.

2.4 Research Gap

While it is clear there are numerous research studies completed around Privacy and Technology for society there seems to be a lack of sufficient research undertaken in AI and Privacy for people living Intellectual Disabilities.

While Rai (2023) highlights the urgent need to look into Ethical and Privacy concerns in AI as a broad societal approach, there is a gap where IDD is concerned and the specific privacy concerns in that group of Society whose needs and wants are more extreme.

By addressing this research gap, we can ensure the conversation includes all vulnerable groups in society in the privacy discussion.

Methodology

3.1 Study Overview

The main goal of this study is to establish how people feel about using Artificial Intelligence in relation to privacy concerns and how Artificial Intelligence can improve Quality of Life for individuals with IDD.

As such a Quantitative Research Strategy was applied and a Descriptive Research approach was adopted to gather our own primary data with the largest number of respondents possible in an efficient manner rather than using smaller focus groups and interviews. According to Ghanad (2023), a survey is the most beneficial way to gather information on attitudes and behaviours of a population.

The Methodology section outlines the Survey Design, Sampling Techniques, Data Collection and Data Analysis methods applied throughout the research to test reliability and accuracy of the data.

3.2 Survey Design

The survey consisted of 8 multiple choice questions followed by 60 questions with a 5-point Likert Scale, described by Joshi et al (2015) as the most effective way to measure human attitudes in social sciences. The aim of the survey was to gather a minimum of 100 responses from people who had a relative or family member living with an Intellectual, Developmental or other Learning Disability. The questions were developed based on personal experience of this area and having discussions with other members of the IDD community as a form of pre-testing the topics.

3.3 Sampling Techniques

The target population for this survey are relatives and family members of an individual living with an IDD. To reach this population a sample was taken from Down Syndrome Limerick that covers numerous age groups, genders, educational achievements and relationships to ensure a diverse range of respondents within the sample.

3.4 Data Collection

The survey was conducted online using University of Galways Microsoft Forms account released on May 13^th for a period of one month. Before release, the required ethics documentation and formal approval to proceed was received from Dr Pierangelo Rosati.

Respondents were required to be over 18 and the survey would close if someone confirmed they were under 18. Respondents were not given a time limit for completion as the survey was quite detailed.

Down Syndrome Limerick and the T21 Journey shared the survey with their members via their websites and through their social media platforms. A total of 63 people completed the survey but as some failed qualifier questions only 58 surveys were valid and used in the analysis.

3.5 Data Analysis

The data collected was exported from MS Forms to MS Excel. A Master file copy was saved, and a file named “SPSS ready” was created for loading and analysis. The statistical software SPSS from IBM was used for the analysis.

Ali et al (2016), discusses the importance of using both Descriptive statistics and Inferential statistics on the data to ensure that meaningful interpretations can be drawn from the data giving something tangible to rows and columns of numbers.

The Descriptive Statistics used to include the mean, median, mode and standard deviation were calculated via SPSS to summarize the data and create data visualisations.

Inferential Statistics such as Correlation and Regression analysis were used to test relationships between different variable and to test the overall Hypothesis.

A Cronbach’s Alpha and factor analysis were used to test the validity and reliability of the collected data.

3.6 Ethical Considerations

From initial conversations with people who have a relationship with someone living with an IDD it was very clear that they are very private about Disability and their family members which is understandable and ensuring that attitudes and beliefs impacting privacy while ensuring anonymity was a key consideration for the research gathering.

DR Pierangelo Rosati reviewed the survey and Ethics forms supplied for the research and signed off the approach. A cover letter in every survey provide participants with detailed information on how the data would be gathered and stored. The surveys’ purpose and was clearly identified and respondents were asked to confirm they consented to proceed with the survey as their choice.

Data Analysis & Results

4.1 Data Overview

The survey was input into SPSS and a number of analysis were carried out to extrapolate some useful information. The below table in fig 2 of descriptive statistics gives some high level details around the data.

The Gender data had 58 valid responses with a high Kurtosis and positive skewness indicating the data was well dispersed tailing off to the right. The Mean indicates more respondents were female.

The Marital Status Data also had 58 valid responses with a sharp Kurtosis and a positive skewness to the right. The Mean indicates that the majority of respondents were married.

The highest level of education achieved showed via the mean as an Undergraduate degree, the Kurtosis was a slight peak and there was a negative Skewness showing the distribution of data tails off to the left.

Figure 2. Descriptive Statistical Analysis

Fig 3 below shows the split by Gender with 74% of repondents classed as female, 24.1% male and 1.7% classed as non binary.

Figure 3. Gender Statistics

Figure 4 below shows that 70.7% of the respondents were married

Figure 4. Marital Status Statistics

Figure 5 below shows that 47% achieved a highest educational achievement at undergraduate degree level

Figure 5. Educational Achievement Statistics

4.2 Tests of Normality

Figure 6 below shows the Kolmogorov-Smirnov and Shapiro-Wilk tests that were conducted to see if the distribution of the Total Privacy variable deviated significantly from a normal distribution.

The Kolmogorov-Smirnov Test returned the below statistical details of note

Statistical Test: The test statistic value is .080.

Significance: The p-value (Sig.) is .200.

The p-value is greater than .05, so we must fail to reject the null hypothesis. This means that there is no significant deviation from normality for the Total Privacy variable based on this test and its results.

The Shapiro-Wilk Test returned the below statistical details of note

Statistical Test: The test statistic value is .972.

Significance: The p-value (Sig.) is .201.

The p-value is greater than .05, so we must fail to reject the null hypothesis. This means that there is no significant deviation from normality for the Total Privacy variable based on this test and its results.

Figure 6 Tests of Normality

Figure 7 below shows the Total Privacy Histogram data was evenly distributed.

Figure 7 Total Privacy Histogram

Figure 8 below shows the Q-Q plot for Total Privacy follows a normal distribution, as most of the data points lie close to the line. The deviations at the end or tails are minor and typical for real-world data. This visual inspection supports the results from the Kolmogorov-Smirnov and Shapiro-Wilk tests, which indicated that the distribution does not significantly deviate from the previous tests of normality.

Figure 8 Total Privacy Q-Q Plot

The survey section on assessing privacy and ethical concerns around AI was for most of the questions strongly negative as represented by figure 9 below. 80% of respondents felt that their data was not safe when stored on mobile devices while 72% of respondents do not trust Organisations to use AI in an ethical manner. A further 38% of repondents do not feel comfortable sharing their data around a family member with IDD on platforms or AI devices.

Figure 9 Assessing Privacy Concerns

4.3 Reliability Testing

Statistical test: Cronbach’s Alpha which is a measure of the values in a scale to ensure they are all correlated and recorded consistently.

Results: The Cronbach’s Alpha value in this analysis for Total Privacy Concerns which can be seen in the Reliability Statistics Fig 1 below  is 0.804, which shows an acceptable value for Cronbach’s Alpha as it is greater than 0.7.

Fig xx Cronbachs Alpha = 0.804

Fig xx Cronbach’s Alpha if item deleted correlation & Positive Corrected item – Total Correlation

Limitations

This research paper like all academic research papers must acknowledge the potential limitations, inaccuracies and risks contained withing the data collected from the respondents via the survey.

Firstly, the most important issue to flag is that the number of expected responses is far lower than the response rate that was anticipated before releasing the survey. However, this is because of a technical issue many respondents encountered while trying to complete the survey where they were blocked by Microsoft from accessing and completing the survey, this was reported by numerous people. See fig xxxx below as a sample of the issue many respondents encountered even though security and access settings for the survey were correctly setup.

Secondly, it is important to highlight that the data collected is based on respondents providing accurate and honest feedback which is outside of the control of the researcher and a potential limitation.

Thirdly, while respondents may have answered questions to the best of their ability it is possible that they did not understand the topic of Artificial Intelligence or how prevalent it is in their daily activities fully. This can be seen in the “Assessing Knowledge” section of our survey.

.

Fourthly, due to the group of respondents that was specifically targeted for this area of research it most be noted that there is a chance of sample bias which is not reflective of wider society.

Finaly the survey depending on the readers viewpoint may have design limitations due to its length and the time required to complete, average completion time was 11 minutes which may indicate a level of respondent fatigue which was confirmed from some feedback received. Also, the format of the questions may contain inherent bias and some leading questions.

Future Suggested Research

The survey results clearly indicate that the respondents have significant concerns around the ethical and privacy issues surrounding the use of AI for family members or relatives living with IDD, yet the respondents also believe in the untapped potential to improve quality of life outcomes in many different aspects of life such as education, communication skills and social skills, therefore future research can explore many different opportunities.

One area that requires further research based on our survey results is new strand of research into the Privacy Calculus Theory within the context of AI for individuals living with an IDD. This theory concludes that individuals weigh the perceived benefits of data sharing against the potential privacy risks, but how should this be addressed for people who sometimes may not have the mental capacity to make informed decisions.

As Rai (2023) pointed out there is an urgent need for Future research which should focus on developing and validating frameworks that balance privacy concerns with the benefits of AI.

Studies should examine the specific types of AI applications mentioned throughout the study for individuals with IDD and their carers that they find most beneficial such as Speech Sounds and Sign Language applications which are clearly high reward and the conditions under which carers are willing to accept privacy risks.

Further longitudinal studies should be undertaken that track the long-term impacts of AI interventions on the quality of life, assessing how privacy concerns evolve over time for individuals and families or relatives of people living with an IDD become more familiar with these technologies.

Research should also explore the role of transparent data governance policies and enhanced consent mechanisms in mitigating privacy concerns. Enforced white box applications for vulnerable groups of society is an area that national governments could fund for the greater good of society.

Understanding how to effectively communicate the benefits and risks of AI effectively to advocacy groups, educational bodies and state agencies can help to develop and create best practices for developers and policymakers, which will ultimately foster a more ethical and user-centered approach to AI implementation.

The key aspect of all future research is that it focuses on a lived experience for people living with an IDD and includes Privacy, People, Process and Technology concerns.

Conclusion

While Artificial Intelligence is evolving continuously it is important that Privacy Concerns and Ethical Considerations are at the heart of all developments especially when we look at vulnerable groups in society such as people living with Intellectual and Developmental Disabilities.

Our study clearly shows that Privacy Calculus Theory is underpinning peoples decisions when it comes to the use of AI driven devices, tools and applications. People are concerned about sharing, storing and engaging with AI but on the other hand there is a clear understanding of how the technology can help people with IDD’s to enhance learning and developmental.

The most heartening take away from this study is that 80% of respondents believed that AI would help their loved ones achieve a better quality of life, which fully deserves to be explored further and developed into a core part of Computational Social Science.

Acknowledgements

I would like to thank Dr Pierangelo Rosati for his feedback and guidance as my supervisor for this research project.

The survey responses would not have been possible with the assistance of Down Syndrome Limerick who shared the survey across their social media accounts.

I would also like to thank The T21 Journey for also sharing the survey across their website and social media platforms.

Appendix

Research Ethics Guidelines for students enrolled in the

BIS Postgraduate Project Module

Project Title

Exploring the Influence of Bias, and Ethical Considerations on the Acceptance and Utilization of AI and Machine Learning to achieve better outcomes in Learning, Communication, and Quality of Life for Individuals living with Intellectual and Developmental Disabilities.

Aims of Research

Under this heading, please give an outline of the significance of the proposed project and an explanation of any expected benefits to individuals, organisations and/or the community in general (100-150 words approximately)

There are many mobile device applications that leverage Artificial Intelligence and Machine Learning for therapies, medical tracking and wellness that are available to carers who have a family member living with Down Syndrome, however there appears to be a reluctance for people to engage with these apps.

I would like to investigate what types and categories of mobile applications are available and what the root cause of the reluctance to engage with these AI technologies is.

Proposed Methods

Under this heading, please give an outline of the proposed methodology, including details of how potential participants will be approached, data collection techniques, tasks participants will be asked to do, and the estimated time commitment involved. This section will vary in length depending on how many different research techniques you intend using and how many different groups of participants you intend involving in your study. However, you should be able to summarise your research methods adequately in under 600 words.

The survey will be administered via MS Forms and all data collected will be stored on University of Galway’s OneDrive.

I am a member of several private parent groups on Facebook such as Raising a child with Down Syndrome in Ireland who have over 2,000 members and this group would be willing to allow me to share my survey with all members.

I am also a member of Down Syndrome Limerick and Family Carers Ireland who regularly assist researchers by sharing surveys, focus groups and studies with their members via email and social media to help further research into different aspects of Down Syndrome that impact on members.

I intend to ask these organisations to share the link to my online survey and collect the anonymous answers for statistical analysis which will form the basis of my conclusion and further research suggestions.

I also own a DS advocacy website called The T21 Journey and was hoping to publish the survey there and to our followers on our various social media channels.

I hope to recruit a minimum of 100 participants from the membership of these groups ro complete my survey.

Ethical Implications of My Study and Steps Taken to Protect Participants:

Under this heading, please describe the ethical implications of your research and provide an overview of the various methods you have used to protect participants in your study from risk. This section will vary in length depending on the ethical implications of your study. However, you should be able to summarise these procedures adequately in under 600 words.

As the research deals with Intellectual Disabilities in particular Down Syndrome it is understandable that people would not want to share private and identifiable information, as such the survey is completely anonymous no names or locations are taken as part of the survey to ensure this.

Privacy is of utmost importance to the integrity of the research.

Once you have completed the sections above to your own satisfaction, please sign one copy and submit them to the module coordinator as per the assignment guidelines.

 Please include copies of the following with your form:

• Your informed consent letter(s)
• Where appropriate, a draft of your questionnaire
• Where appropriate, a draft of your interview questions or in the case of open-ended interviews, your topics

Please note that you should not engage in any primary research until your supervisor has contacted you . If you undertake any primary research involving human participants without first submitting a completed research ethics form and assessment by the module coordinator, this research cannot be considered for the final evaluation.

Exploring the Influence of Bias, Privacy and Ethical Considerations on the use of AI and Machine Learning to achieve better outcomes in Learning, Communication, and Quality of Life for Individuals living with Intellectual Developmental Disabilities.

Dear Participant,

As part of my Master of Science in Business Analytics major project, I am conducting research into the area of Exploring the Influence of Bias, Privacy and Ethical Considerations on the Acceptance and Utilization of AI and Machine Learning to achieve better outcomes in Learning, Communication, and Quality of Life for Individuals living with Intellectual and Developmental Disabilities in the University of Galway.

I am investigating this because I want to understand if personal bias and privacy concerns prevent carers from using Artificial Intelligence apps on mobile devices to support their family member living with Down Syndrome in different areas of therapy, health management and education.

I am inviting you to participate in this research project because of your membership and participation in the Down Syndrome Community. Accompanying this letter is a short questionnaire that asks a variety of questions about Artificial Intelligence and privacy concerns.

I am asking you to review the questionnaire and, if you choose to do so, complete it and submit it back to me. It should take you about five minutes to complete. The questionnaire does not require you to give your name or any other information that might identify you.

The survey will be completed via MS Forms and all data will be stored on University of Galway’s OneDrive storage accounts.

Information compiled from the questionnaire will be reported in aggregate form and individuals will remain anonymous. No information you give will be shared with any other individual.

Through your participation I hope to understand how we can tailor new applications to help people living with Down Syndrome to reach their full potential addressing all concerns carers may have.

I hope that the results of this survey will be useful for developing applications, action groups and activities with a view to embracing technology. While I do have the support of Pierangelo Rosati Associate Professor in Digital Business and Society who is my Research Supervisor to engage in this research, it is being conducted by me in a personal capacity.

You do not have to participate in this study if you do not wish to do so. I would like to thank you for taking the time to read this letter.

Regardless of whether you choose to participate, please let me know if you would like a summary of my survey findings.

Do you consent to participate in this survey? YES/No skip logic if no

Are you 18 years old or older?  YES/NO skip logic if no

Key terms Overview

IBM defines Artificial intelligence, or AI, “as a technology that enables computers and machines to simulate human intelligence and problem-solving capabilities.”

IBM defines Machine learning (ML) “as a branch of artificial intelligence (AI) and computer science that focuses on the using data and algorithms to enable AI to imitate the way that humans learn, gradually improving its accuracy.”

Bias is the tendency to show support or opposition towards one idea, person, or thing regardless of the full facts.

Survey

1. Sex: _____Male _____Female (please tick whichever applies)

2. Age: ___________ 18-24 _________25 -34 ________ 35-44

_______45-54 ______55-64______65+

3. What is your marital status: (please tick whichever best describes your current situation) _____single _____in a steady relationship _____living with partner _____married

_____separated _____divorced _____Widowed.

4. Do you care for or have a relative living with Down Syndrome: _____Yes _____No

5. How do you define your role: _______Carer _______Therapist ______Family Member

     _______All of the Above

6. What is the highest level of education that you have completed: (please tick the highest level you have completed).

_____ primary school

_____ secondary school

_____ some additional training (apprenticeship, trades etc.)

 _____ completed undergraduate University programme.

_____ completed postgraduate University programme.

______apprenticeship or trade

Section 1 Assessing knowledge around Artificial Intelligence & Machine Learning

Please read through the following statements and decide how much you either agree or disagree with each.  Using the scale provided write the number that best indicates how you feel on the line next to each statement. 

strongly disagree     1        2       3        4        5        strongly agree.

1. I understand the basic concepts of Artificial Intelligence (AI)
2. I understand that computers, applications, and programs can learn from data.
3. I feel I could explain AI to someone who has never heard of it before.
4. I have used online Virtual assistants, chat bots and smart home devices.
5. I use apps such as Netflix, YouTube, Twitter, and other social media apps which are based on recommendation system algorithms based on my behaviours.
6. I am confident in my ability to recognise AI technologies in my daily activities.
7. I believe I can distinguish between factual and untrustworthy claims regarding AI.
8. I believe I can identify tasks that would be suitable for AI use cases.
9. I believe AI will be instrumental in future products and services.
10. I understand the difference between Artificial Intelligence and Human Intelligence

Section 2 Assessing Bias, Privacy & Ethical Concerns

Please read through the following statements and decide how much you either agree or disagree with each.  Using the scale provided write the number that best indicates how you feel on the line next to each statement. 

strongly disagree     1        2       3        4        5        strongly agree.

1. I am concerned about the potential bias, accuracy, and fairness regarding disability used in the training of artificial intelligence models.
2. I do not believe enough is done to safeguard people’s data privacy online.
3. I believe that AI models can be trained and designed to remove bias.
4. I am open minded around AI usage and products.
5. Concerns around my sibling or child’s data regarding their disability stops me from using AI technologies for therapy, education, and communication.
6. AI and its further integration in society concerns me.
7. My personal details are safe when stored on mobile apps.
8. Two Factor Authentication ensures my data is safe.
9. I believe the AI tech I use should explain how it makes its decisions.
10. I trust organisations that use AI to behave in an ethical way.
11. Regardless of how convenient the product or service is I do not trust AI tech as it will impact on my privacy.
12. AI developers are obliged to ensure that their products are safe for people with Intellectual disabilities to use.
13. AI based healthcare and learning tools compromise the privacy of people with intellectual disabilities.
14. Lack of understanding stops the introduction of AI tools for the Intellectual Disability community.
15. Caregivers are ready to accept AI technology assistance to achieve better outcomes.

Section 3 Improved Learning Outcomes using Artificial Intelligence

Please read through the following statements and decide how much you either agree or disagree with each.  Using the scale provided write the number that best indicates how you feel on the line next to each statement. 

strongly disagree     1        2       3        4        5        strongly agree.

1. There are no doubt AI technologies can improve learning outcomes for children with intellectual disabilities.
2. AI tools remove the emotion and make the learning experience more efficient.
3. AI tools can adapt and tailor the teaching based on the child’s abilities after initial screening tests.
4. AI tools can engage children’s sensory needs in ways human teaching cannot achieve through changing tactile, auditory and visual learning.
5. Children will be more motivated to learn based on the Apps ability to learn their behaviours and tailor the sessions.
6. Cost is not a barrier to usage of AI learning applications.
7. AI tools and technologies can make progression through the school years easier as the child will have a familiar tool.
8. I believe AI can help children and adults achieve better learning outcomes when used with traditional teaching methods.
9. AI technology is a crucial tool for the Department of Education to invest in for children with Intellectual Disabilities
10. AI can foster an inclusive teaching environment.

Section 4 Improved Communication Outcomes using Artificial Intelligence

Please read through the following statements and decide how much you either agree or disagree with each.  Using the scale provided write the number that best indicates how you feel on the line next to each statement. 

strongly disagree     1        2       3        4        5        strongly agree.

1. AI can ensure better communication skills for individuals with Intellectual Disabilities
2. AI tools are much more suitable for teaching sign language.
3. Speech and sound applications can improve language skills.
4. Devices with AI communication abilities can help people living with Intellectual disabilities become more confident.
5. AI communication devices are easily available.
6. AI can help remove barriers to communication for people with intellectual disabilities.
7. AI communication tools will lead to more independent living for people with intellectual disabilities.
8. Speech and Language therapy can be transformed with AI technologies.
9. AI tools can help improve literacy and numeracy skills.
10. AI tools will help people living with intellectual disabilities become more emotionally aware and expressive.

Section 5 Assessing improved Quality of Life Opportunities

Please read through the following statements and decide how much you either agree or disagree with each.  Using the scale provided write the number that best indicates how you feel on the line next to each statement. 

strongly disagree     1        2       3        4        5        strongly agree.

1. AI Technologies can ensure a better independent living opportunity.
2. I believe AI tools can help improves the safety and security of people with intellectual disabilities.
3. AI tools can help with building life skills for people with intellectual disabilities.
4. I believe AI tools can help with nutrition and dietary choices for people with intellectual disabilities.
5. I would use a tool that stored all my child’s progress, therapies, appointments, medical history, and medication all in one place that facilitated analytics.
6.  AI tools will further foster inclusion in society.
7. I believe AI tools will help my child develop a personalised routine that helps them build independent living skills.
8. AI technologies will enable more people with Intellectual disabilities participate in the workforce.
9. Caregivers and healthcare professionals will be able to provide better medical outcomes for people with intellectual disabilities by using predictive AI technologies.
10. Thinking about my child’s needs, I believe AI can improve their quality of life.
11. The potential for future AI driven technologies to help people with intellectual disabilities excites me.
12. Society is aware of the potential to help people with intellectual disabilities using technology.
13. AI technology can help create a more balanced environment for care givers helping free up more personal time for carers.
14. There will be an increase in people with intellectual disabilities going to third level education due to assistance from AI.
15. I believe data analytics gathered from the technologies my child engages with shared with therapists will improve all areas of development.

Fig xx Total Privacy Concerns Case Processing Summary

Fig xx Total Privacy Concerns Descriptives

Fig xx Extreme Values

Fig xx Total Privacy Concerns Case Processing Summary

Fig xx Total Privacy Concerns Descriptives

Fig xx Extreme Values

References

Aggarwal, G., Singh, L. (2018) ‘Evaluation of Supervised Learning Algorithms Based on Speech Features as Predictors to the Diagnosis of Mild to Moderate Intellectual Disability’  3D Res 9, 55, available: https://doi-org.nuigalway.idm.oclc.org/10.1007/s13319-018-0207-6

Ali, Z., & Bhaskar, S. B. (2016). Basic statistical tools in research and data analysis. Indian journal of anaesthesia, 60(9), 662–669. https://doi.org/10.4103/0019-5049.190623

Broda MD, Bogenschutz M, Dinora P, Prohn SM, Lineberry S, Ross E. (2021) ‘Using machine learning to predict patterns of employment and day program participation’, AJMR. American Journal on Intellectual and Developmental Disabilities. 2021;126(6):477-491. Available: https://nuigalway.idm.oclc.org/login?url=https://www.proquest.com/scholarly-journals/using-machine-learning-predict-patterns/docview/2586263300/se-2. doi: https://doi.org/10.1352/1944-7558-126.6.477.

Culnan, M. J., & Armstrong, P. K. (1999). Information Privacy Concerns, Procedural Fairness, and Impersonal Trust: An Empirical Investigation. Organization Science, 10(1), 104–115. http://www.jstor.org/stable/2640390

Fox, G., Clohessy, T., Van der Werff, L., Rosati, P., & Lynn, T. (2021). Exploring the competing influences of privacy concerns and positive beliefs on citizen acceptance of contact tracing mobile applications. Computers in Human Behavior, 121, 106806. https://doi.org/10.1016/j.chb.2021.106806

Gaurav Kumar Yadav, Benigno Moreno Vidales, Hatem A Rashwan, Joan Oliver, Domenec Puig, G.C. Nandi, Mohamed Abdel-Nasser, (2023), ‘Effective ML-based quality of life prediction approach for dependent people in guardianship entities’, Alexandria Engineering Journal, V65, 909-919,

ISSN 1110-0168, https://doi.org/10.1016/j.aej.2022.10.028,

Available:  https://www-sciencedirect-com.nuigalway.idm.oclc.org/science/article/pii/S1110016822006846?via%3Dihub

Ghanad, Anahita. (2023). An Overview of Quantitative Research Methods. INTERNATIONAL JOURNAL OF MULTIDISCIPLINARY RESEARCH AND ANALYSIS. 06. 10.47191/ijmra/v6-i8-52.

Gupta, C., Chandrashekar, P., Jin, T. et al. (2022) ‘Bringing machine learning to research on intellectual and developmental disabilities: taking inspiration from neurological diseases’, Journal of Neurodevelopment Disorders, 14, 28, available: https://doi.org/10.1186/s11689-022-09438-w

Haug, Maximilian; Rössler, Philipp; and Gewald, Heiko, “HOW USERS PERCEIVE PRIVACY AND SECURITY RISKS CONCERNING SMART SPEAKERS” (2020). In Proceedings of the 28th European Conference on Information Systems (ECIS), An Online AIS Conference, June 15-17, 2020.
https://aisel.aisnet.org/ecis2020_rp/129

Heins, M., Johnson, R., & Smith, P. (2020). “AI in Personalized Learning for Children with Autism.” Journal of Autism and Developmental Disorders, 50(11), 3927-3941.

Hutchinson Ben, Vinodkumar Prabhakaran, Emily Denton, Kellie Webster, Yu Zhong, and Stephen Denuyl. 2020. Unintended machine learning biases as social barriers for persons with disabilitiess. SIGACCESS Access. Comput., 125, Article 9 (October 2019), 1 pages. https://doi.org/10.1145/3386296.3386305

Johnson, R., Garcia, L., & Smith, P. (2019). “AI-driven Health Monitoring Systems for Individuals with IDD.” Journal of Intellectual Disability Research, 63(8), 789-799.

Jones, K., Zahrah, F., & Nurse, J. R. (2024). Embedding Privacy in Computational Social Science and Artificial Intelligence Research. ArXiv. https://doi.org/10.36190/2024.18

Joshi, Ankur & Kale, Saket & Chandel, Satish & Pal, Dinesh. (2015). Likert Scale: Explored and Explained. British Journal of Applied Science & Technology. 7. 396-403. 10.9734/BJAST/2015/14975.

Krasnova, Hanna; Veltri, Natasha F.; and Günther, Oliver (2012) “Self-disclosure and Privacy Calculus on Social Networking Sites: The Role of Culture – Intercultural Dynamics of Privacy Calculus,” Business & Information Systems Engineering: Vol. 4: Iss. 3, 127-135.
Available at: https://aisel.aisnet.org/bise/vol4/iss3/4

Maram Fahaad Almufareh, Samabia Tehsin and Mamoona Humayun et al. Intellectual Disability and Technology: An Artificial Intelligence Perspective and Framework. JDR. 2023. Vol. 2(4):58-70. DOI: 10.57197/JDR-2023-0055 https://www.scienceopen.com/hosted-document?doi=10.57197/JDR-2023-0055

NIHR (2023) National Institute for Health and Care Research announces £10m funding to artificial intelligence research for multiple long term conditions, available: https://www.nihr.ac.uk/news/more-than-10m-awarded-to-artificial-intelligence-research-for-multiple-long-term-conditions/31373 (accessed 13 Nov 2023)

Princi, E., & Krämer, N. C. (2020). Out of Control – Privacy Calculus and the Effect of Perceived Control and Moral Considerations on the Usage of IoT Healthcare Devices. Frontiers in psychology, 11, 582054. https://doi.org/10.3389/fpsyg.2020.582054

Rai, Paras. (2023). Ethics in AI: A Deep Dive into Privacy Concerns. Available at https://www.researchgate.net/publication/376518039_Ethics_in_AI_A_Deep_Dive_into_Privacy_Concerns

Song C, Jiang ZQ, Hu LF, Li WH, Liu XL, Wang YY, Jin WY, Zhu ZW, (2022) ‘A machine learning-based diagnostic model for children with autism spectrum disorders complicated with intellectual disability’, Frontiers in Psychiatry. 2022 Sep 21;13:993077. doi: 10.3389/fpsyt.2022.993077. PMID: 36213933; PMCID: PMC9533131. Available: https://www-ncbi-nlm-nih-gov.nuigalway.idm.oclc.org/pmc/articles/PMC9533131/

Standen, Penelope J, David J Brown, Mohammad Taheri, Maria J Galvez Trigo, Helen Boulton, Andrew Burton, Madeline J Hallewell, James G Lathe, Nicholas Shopland, Maria A Blanco Gonzalez, Gosia M Kwiatkowska, Elena Milli, Stefano Cobello, Annaleda Mazzucato, Marco Traversi, and Enrique Hortal, (2020), “An Evaluation of an Adaptive Learning System Based on Multimodal Affect Recognition for Learners with Intellectual Disabilities.” British Journal of Educational Technology 51.5: 1748-765. Web. Available: https://bera-journals-onlinelibrary-wiley-com.nuigalway.idm.oclc.org/doi/full/10.1111/bjet.13010

Trewin Shari, Sara Basson, Michael Muller, Stacy Branham, Jutta Treviranus, Daniel Gruen, Daniel Hebert, Natalia Lyckowski, and Erich Manser. 2019. Considerations for AI fairness for people with disabilities. AI Matters 5, 3 (September 2019), 40–63. https://doi.org/10.1145/3362077.3362086

Wang, T., et al. (2016). “Intention to disclose personal information via mobile applications: A privacy calculus perspective.” International Journal of Information Management 36(4): 531-542. https://doi.org/10.1016/ j.ijinfomgt.2016.03.003. Publisher version of record available at: https://www.sciencedirect.com/science/article/pii/ S0268401215300797

Whelan, E., Lang, M. and Butler, M. (2024), “Beyond lazy; external locus of control as an alternative explanation for the privacy paradox”, Internet Research, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/INTR-04-2023-0282