AI Education Framework for Analysis and Profiling of Students Using Machine Learning
- B.Venura Lakshman
- Dr J.V. Wijayakulasooriya
- Prof T.M.S.S.K. Yatigammana
- 462-469
- Dec 12, 2024
- Education
AI Education Framework for Analysis and Profiling of Students Using Machine Learning
B.Venura Lakshman1, Dr J.V. Wijayakulasooriya2, Prof T.M.S.S.K. Yatigammana3
1Bhiksu University of Sri Lanka, Anuradhapura
2,3University of Peradeniya, Peradeniya
DOI: https://doi.org/10.51584/IJRIAS.2024.911035
Received: 25 October 2024; Accepted: 06 November 2024; Published: 12 December 2024
ABSTRACT
Today, in many educational institutes use AI technologies, to enhance the educational level of the institute. Using such technologies, the institute can get many benefits like identify weak students earlier, teach technology easily…etc. Also, this can be used to select a proper AI-based Electronic learning system to the institute. In addition to that, by adopting AI to an educational institute academic staff also can get many benefits like how to change teaching method and selecting appropriate teaching method…etc. To adopt AI to an educational institute, it is required to follow a comprehensive and methodological manner. For this reason, a framework is required to adopt AI to an educational institute. But there is no such comprehensive framework to adopt AI to an educational institute. So, education institutes feel a gap between their current position as an AI-enabled institutes. The following framework is introduced to cover up this gap.
Keywords: AI, Educational Institute, E Learning, framework, Machine Learning
METHODOLOGY
This conceptual framework is designed for the transformation a normal educational institute into an AI enable educational institute. The framework consists of the following components. The framework can be tested and utilized at various levels according to needs and requirements of the educational institute.
Core – The nucleus of the Framework, which is a taxonomy of high-level AI-based education outcomes that can help any educational organization manage its education activities. The framework core components are a hierarchy of Functions, Categories, and Subcategories that detail each outcome. These outcomes can be understood by a broad audience, including teachers, educational managers, students and parents, regardless of their AI expertise. Because the outcomes are size of institute, country, and technology-neutral, they provide an organization with the flexibility needed to address its unique education requirements, technologies, and mission considerations.
Organizational profiles – Organization and other personal profiles which consists of various data about objectives and learning and teaching particulars…etc.
Tiers – Which can be applied to CSF Organizational Profiles to characterize the rigor of an organization’s cybersecurity risk governance and management practices. Tiers can also provide context for how an organization views cybersecurity risks and the processes in place to manage those risks.
Safety and wellbeing – All categories and sub categories to take the safety and wellbeing of all parties of the teaching and learning process.
Tools – All AI based tools which are appropriate for different levels of the teaching, learning and management process.
INTRODUCTION TO THE CORE OF THE FRAMEWORK
A set of cybersecurity outcomes arranged by Function, then Category, and finally Subcategory, as depicted in Fig. 1.
Fig 01
Introduction to Organizational profiles
This describes an organization’s current profiles which describe the situation of the person and the organization. In addition to that, this includes targets of each. Following actions should be implemented for profiles.
I. Scope the organizational profiles.
II. Gather required particulars.
III. Add particulars.
IV. Analyze the gaps and create action plans for each profile.
V. Implement the action plan and update the profiles.
Tiers
In this section, it is shown the levels of implementation of the framework. There are four levels to implement in any educational institute. (Fig.02)
Fig.02
Fig.03
Safety and wellbeing
This component deals with safety and wellbeing. It is used to benefit all members of the school community. All tools and resources should be utilized in ways that do not harm the wellbeing and safety of any member of the education institute’s community while protecting human rights.
In addition to that, transparency of the institute’s educational activities should be kept. Teachers, students, staff, parents and education managers have access to clear and appropriate information and guidance about AI tools. Also, robust cyber-security measures are implemented to protect the integrity, safety and availability of the institute’s infrastructure, AI tools, and associated data.
AI tools
These tools are used to analyze and profile students using machine learning. So, institute authorities can identify at-risk students earlier and utilize suitable precautions for the students. In addition to that, extra tools can be utilized to maintain or upgrade the level of students.
Implementing the AI framework
Implementing any AI framework in an educational institution brings exciting opportunities but also presents several potential obstacles and challenges. These obstacles and challenges often revolve around data privacy concerns, budgetary constraints, and the need for comprehensive staff training. [1] In this section there is a discussion of these challenges and suggestions for overcoming them.
Data Privacy Issues
Challenge- Educational institutions handle a vast amount of sensitive data, including personal information about students, teachers, and staff. The implementation of AI systems often requires access to large datasets for training models (e.g., grades, attendance, behavior records), which can raise concerns about data privacy and security. AI systems might also process this data in ways that may not align with privacy laws like Data Protection Act
Suggestions for Overcoming-
• Data Anonymization- Use techniques to anonymize personal data before feeding it into AI systems. This reduces the risk of violating privacy and helps protect sensitive information.
• Compliance with Legal Frameworks- Ensure the AI framework adheres to regional and international data protection laws (e.g., In Sri Lanka, Data protection Act…etc.). In addition to that, Institutions should appoint a Data Protection Officer (DPO) to oversee compliance.
• Transparent Data Practices- Inform stakeholders (students, staff, parents) about the AI system’s data usage and obtain explicit consent. Transparency helps in building trust and managing concerns around data privacy.
• Implement Robust Security Measures- Use encryption, secure data storage, and access controls to protect the data used in AI systems.
Budgetary Constraints
Challenge- Implementing AI systems can be expensive. Costs may include hardware, software, personnel, training, data charges and ongoing maintenance. For many educational institutions, especially public schools or small colleges, these budgetary limitations can make it difficult to adopt cutting-edge technology.
Suggestions for Overcoming-
• Phased Implementation- Instead of a full-scale roll-out, adopt a phased approach where the AI system is implemented gradually. Start with adoptable tiers (e.g., AI-driven grading tools, chatbot assistance) and scale as the budget allows.
• Seek External Funding- Explore government grants, private sector partnerships, or funding from educational technology organizations. Many companies are willing to partner with educational institutions for AI adoption as part of their corporate social responsibility (CSR).
• Leverage Open-Source Solutions and supports- Many open-source AI platforms and tools can help reduce initial costs.
• Collaborate with Industry Partners- Form partnerships with tech companies or universities that can provide both financial support and technical expertise.
Need for Staff Training
Challenge – AI adoption in education requires educators and administrative staff to possess a certain level of technical knowledge and skills to operate and manage AI systems effectively. Many teachers may not be familiar with AI technologies, and there may be resistance to change due to the learning curve associated with new tools. [5]
Suggestions for Overcoming-
• Continuous Professional Development- Implement ongoing training programs for staff. This can include both technical training (e.g., understanding how AI models work) and practical training (e.g., using AI tools in the classroom).
• Foster a Growth Mindset- Encourage a culture of lifelong learning where staff are supported in their transition to AI-driven practices. Engage teachers in the process by demonstrating how AI can enhance their teaching methods rather than replace them.
• Collaborative Training Models- Partner with universities or AI companies to provide expert-led workshops and courses. Alternatively, consider remote learning options (MOOCs, webinars) that staff can attend at their convenience.
• Peer Mentoring- Encourage staff who are more familiar with AI tools to mentor colleagues, fostering collaboration and knowledge sharing across the institution.
Cultural and Organizational Resistance
Challenge- Resistance to AI adoption can stem from staff, students, and even parents who may be apprehensive about the changes AI will bring to the educational environment. Concerns may include fears of technology replacing jobs, a lack of understanding about AI, or a belief that AI systems may be biased or unfair.
Suggestions for Overcoming-
• Change Management- Implement a well-structured change management process that involves all stakeholders from the outset. Clearly communicate the benefits of AI in improving educational outcomes and reducing administrative burdens.
• Involve Stakeholders in Decision-Making: Engage teachers, students, and parents early in the decision-making process. Allowing them to participate in discussions about AI adoption can alleviate concerns and promote buy-in.
• Ethical AI Use- Develop and enforce ethical guidelines for AI use in education, ensuring that systems are transparent, non-biased, and equitable. Regular audits of AI systems can help identify and address potential issues early.
AI Bias and Fairness
Challenge- AI systems, particularly those trained on large datasets, can inherit biases present in the data. In educational settings, this could result in biased grading, assessments, or resource allocation, which can perpetuate inequalities.
Suggestions for Overcoming-
• Bias Audits and Testing- Regularly audit AI models for biases and fairness. This involves testing the system on diverse datasets and ensuring that it does not favor one group of students over others.
• Diversified Data Sets- Ensure that the data used to train AI systems is diverse, representative, and free from bias. This will help the AI model make fair decisions that do not disadvantage any group.
• Transparent Algorithms- Choose AI models that are interpretable and explainable so that educators and administrators can understand how decisions are being made. This transparency can help address concerns and build trust in the system.
Technical Infrastructure and Support
Challenge- AI systems require a robust technical infrastructure, including sufficient computing power, internet connectivity, and data storage capabilities. Many educational institutions, especially those in rural or low-income areas, may lack the necessary resources to support such infrastructure.
Suggestions for Overcoming-
• Cloud-Based Solutions- Utilize cloud-based AI platforms that reduce the need for local computing power and storage. Cloud services often offer scalable resources that can grow with the institution’s needs.
• Infrastructure Partnerships- Seek partnerships with tech companies to upgrade or provide the necessary infrastructure. These companies may offer discounts or donations to educational institutions.
• Focus on Low-Cost AI Solutions- Consider AI solutions that are specifically designed to be resource-light or that can run on existing infrastructure, such as lightweight chatbots or any other simpler AI tools.
Protecting Student and Staff Data
Data privacy and security should be a priority to safeguard sensitive information in educational environments. Following steps will address how to protect data.
Implement Data Minimization
Collect only the data necessary for the educational purpose at hand, and avoid storing excessive or irrelevant information.
E.g. If AI is used for personalized learning, only collect the data required to analyze student performance and learning patterns, such as grades or engagement metrics, without gathering unnecessary personal details like home addresses or phone numbers.
Enforce Data Encryption
Ensure that all sensitive data is encrypted both in transit (while being sent over networks) and at rest (while stored on servers or databases).
E.g. Use encryption protocols like TLS (Transport Layer Security) for online platforms and AES (Advanced Encryption Standard) for database storage.
Provide Access Control and Authentication
Limit access to sensitive data based on roles (e.g., teachers, administrators) and ensure proper authentication (e.g., multi-factor authentication) for accessing such data.
E.g. Only authorized staff members can access detailed student performance records, and login attempts are protected by two-factor authentication (2FA) to prevent unauthorized access.
Adhere to Legal Frameworks
Follow relevant privacy laws and regulations, such as FERPA (Family Educational Rights and Privacy Act) in the U.S. or GDPR (General Data Protection Regulation) in the EU, Data protection Act in Sri Lanka to ensure that data is handled properly.
E.g. Implement policies that allow students and parents to request access to or deletion of their data, as required by these regulations.
Ensuring Transparency
Transparency in how AI tools and systems are used in educational settings is crucial to maintaining trust among students, parents, and staff. Here’s how transparency can be ensured. Also transparency is a level of quality in an educational institute.
Explain AI Models and Their Purpose
Clearly communicate how AI models work, what data they use, and their intended educational benefits to students, teachers, and parents.
E.g. If an AI system is used to recommend learning materials based on student behavior, provide a plain-language explanation of how the AI selects content and what data it considers.
Offer Clear Opt-In and Opt-Out Options
Allow individuals to voluntarily choose whether their data will be used in AI-driven platforms, with an easy process to opt-out or revoke consent.
E.g. A school can inform parents and students about data collection practices for AI-powered assessment tools, with the option to opt out if they do not want their data used for analysis.
Publish AI Impact Assessments
Conduct and publish impact assessments to evaluate potential risks or biases that the AI system might introduce, such as reinforcing gender or racial disparities.
E.g. A school might publish an assessment on how its AI-powered grading system has been tested for fairness, outlining steps taken to prevent biased evaluations.
Provide Clear Feedback and Error Reporting Mechanisms
Ensure students and staff can easily report any issues or mistakes that arise from AI tools, and provide feedback channels to improve the systems.
E.g. A learning management system (LMS) that uses AI to recommend study materials should include a feature where students can flag irrelevant or inaccurate recommendations.
Maintaining Ethical AI Usage
Ethical AI usage is critical to prevent harm, bias, and discrimination. Here are steps to uphold ethical principles to maintain ethical AI usage.
Implement Bias Detection and Mitigation
Regularly test AI systems for bias in decision-making and take corrective actions when bias is detected. [4]
E.g. Use diverse training data when developing AI models to avoid reinforcing societal biases, and regularly audit the system to check for disparities in outcomes based on gender, race, or socioeconomic status.
Design AI to Enhance Educational Equity
Ensure that AI systems are designed to support and improve learning opportunities for all students, especially those from underserved or vulnerable groups.
E.g. AI-powered tools for personalized learning should not favor certain groups of students (e.g., higher-income students) but rather aim to support struggling learners with tailored content.
Promote Human Oversight and Decision-Making
Ensure that AI systems do not replace human educators but rather support them. Teachers should have the final say in critical decisions, such as student assessments.
E.g. An AI grading system could be used for initial assessments, but teachers would review the results to ensure they align with students’ abilities and any special needs.
Safeguard Against Over-automation
Avoid the over-reliance on AI systems, which can lead to disengagement or neglect of individual needs.
E.g. Rather than using AI exclusively to monitor student behavior or learning progress, schools should incorporate human interactions where necessary to ensure the holistic development of students.
Foster Ethical AI Development and Research
Involve ethical committees or review boards in the development of AI systems to ensure that they adhere to ethical guidelines and standards.
E.g. Before deploying AI systems, a school may have an ethics board review the system’s design to check for fairness, transparency, and respect for privacy.
RESULTS AND DISCUSSION
This framework is a basic tool for transforming an educational institute into an AI enabled educational institute. This framework has been developed by various stages of implementation. Many educational institute are suffering from various low tech infrastructure and funding problems. So, they can go various tiers as mentioned in the framework. In addition to that, framework is not limited to certain tools or models. So, users can utilize the framework with any tools and methods.
CONCLUSION
This paper introduces AI framework to adopt AI enable framework to suitable any educational institute. Also, it is discussed challenges and other issues in implementing AI framework in Education.
REFERENCES
- Alaa el-Halees, “Mining students data to analyze e-Learning behavior: A Case Study”, 2009.
- J. Han and M. Kamber, “Data Mining: Concepts and Techniques,” Morgan Kaufmann, 2000.
- Siemens, G., & Gruber, T. R. (2019). Learning analytics: Insights from education and training. Springer Nature.
- Johnson, L., Adams, S., Cummins, M., Estrada, V., Freeman, A., Hall, C., … & Willett, J. (2017). Learning analytics: An emerging field of research and practice. International Society for Technology in Education (ISTE).
- Jiao, Q. G., & Yan, Z. G. (2019). Artificial intelligence in education: A review. International Journal of Artificial Intelligence in Education, 29(1), 147-168.