INTRODUCTION
After Covid 2019 people wanted an answer to the question “How should one continue one’s education while staying in lockdowns and restrictions?”. The accessibility of the Internet made taking classes and lessons much easier. With an increasing young population and demand for education there was a stress in creating content for the sole purpose of education, making lessons, planning courses, creating content, evaluating content etc. Here comes a technology that can imitate (to a certain extent) human intelligence, also known as Artificial Intelligence. A.I. helps augment humans in their tasks and processes and makes them more efficient. But using A.I. also creates many legal challenges that need to be navigated. Let us Discuss.
HOW DOES ARTIFICIAL INTELLIGENCE HELP IN EDUCATION
A.I. helps simulate human intelligence into a computer which helps it think and act like a human. It helps both teachers and students in making the education system better and more effective. A. I can help automate basic activities in education like it is possible to automate the grading for nearly all types of Multiple choice questions and fill in the blanks. Teachers can’t be present with students all the time and each student is not smart enough to grasp all the things at once, with these A.I. programs, students can learn the fundamentals of mathematics, writing, and other subjects. A.I. can also provide feedback to teachers and students as per requirements. With the current education system, it is very difficult to find gaps in learning and also in Industry-academia linkages. A.I. can help bridge this gap and provide necessary feedback whenever needed. A. I also help with personalizing curriculums and determining the needs of every student, with the help of A.I. Teachers can come up with a tailor-made study plan for every student. A.I. tools can be successfully trained to help a group of students with special needs and also it makes education accessible to students worldwide.
ENSURING DATA PRIVACY AND PROTECTION
Although A.I. offers significant benefits, such as personalized learning and automation there are always certain questions that need to be answered regarding the collection and processing of Data especially by A.I. systems. AI in education collects a lot of student data including personal information, academic performance, attendance records, behavior and even spending records. The shift from on-paper to digital methods of data collection has led to increased efficiency and accuracy for academic behaviour and patterns. This extensive collection has led to certain privacy concerns.
The extensive collection of data is analyzed to create a personalized experience for the students and other academicians. The General Data Protection Regulation in the European Union and the Digital Personal Data Protection Bill of India have critical guidelines that regulate the use of Personal data. Getting explicit consent, transparency in Data collection and restricting the scope of usage of data for specific educational purposes. The General Data Protection Regulation ratifies the ‘right to be forgotten’ which allows users to request deletion of their data. The Digital Personal Data Protection Bill gives stress to limiting the use of such data only for Security measures.
DETERMINING THE OWNERSHIP OF AI-GENERATED EDUCATIONAL MATERIALS
There are copyright laws that give ownership credits to human creators. Identifying the owner when an AI model generates content becomes tricky when the generator autonomously generates content. In some domains, the authorship may be adhered to by the organization which operates or programs the A.I. but some may also consider the content to be public due to lack of authorship. This duality creates a need for proper organizational policies and codes to give definitions of A.I. generated educational content.
A.I. models get their training done on large datasets, which can sometimes include copyright material. This becomes the cause for significant legal concerns. Some arguments consider this use as ‘fair use’, but in the majority of the cases, this interpretation is not accepted. Notably several lawsuits have been filed against these A.I. developers for allegedly using copyrighted content for training their models. For example, The New York Times requested a federal judge to make OpenAi ‘identify and admit’ all the newspaper articles that the company has used as a dataset to train. A few days later GEMA, a German licensing body filed a lawsuit against the same OpenAi alleging that OpenAi uses GEMA’s ‘repertoire’ as a dataset to train its system.
POTENTIAL FOR BIASES, LEGAL IMPLICATIONS
A.I. Models learn from pre-created data which may contain certain societal biases. When data with bias is used to train the models then the results might inherently incorporate the biases. For example when a type of data from a population where certain demographic communities were underrepresented the model might continue to screen in such a way that it favours the historically dominant group over the underrepresented ones. Certain discriminatory outcomes come out of these biased A.I. Systems. There are anti-discriminatory laws in certain jurisdictions that prevent practices that give unfair disadvantages to individuals from protected race, gender or disability. If any academic institution employs these practices then they can undergo severe legal complications. A good example would be when an A.I. model in which grade assessment gives lower marks or ranking to a student from a certain ethnic background then this practice can be seen as discriminatory and lead to potential lawsuits.
PRACTICES TO MITIGATE CHALLENGES
A.I. is inevitable as it assists humans to a great extent. But it comes with challenges of privacy, security and ownership of data. To ensure an ethical and effective implementation we must ensure:
Educational and Academic institutions must ensure strict and comprehensive frameworks related to data governance that adhere to data protection laws such as the GDPR in the EU and the Digital Personal Data Protection Bill, of 2023 in India. They must also ensure that privacy is considered during the development processes of the systems which as a result ensures that data collection is minimised and the consent of the user is obtained every time. We must provide training sessions that will equip educators with the knowledge of A.I. tools and their responsible usage. There must be communities of practice where educators using AI platforms can share experiences for integrating AI with education.
Organisations must regularly evaluate A.I. models for any potential biases and adjust accordingly to ensure equity and fairness. During the training process use of diverse and representative datasets must be ensured which would lower the risk of outputs with biases.
CONCLUSION
Tech is evolving. With the fast-paced growth of technology, we cannot ignore A.I., especially its role in transforming how we receive and impart education. From providing personalized learning experiences and automating basic administrative work A.I. becomes a revolution. However, the integration of AI into education comes with a plethora of legal concerns. Privacy of Data, I.P. rights, biases and ‘who would be held responsible’ for decisions made by A.I. models are all questions that need to be answered. To fully capitalize on AI’s capabilities and address its challenges, educational institutions, policymakers, and developers need to work collectively to develop strong legal guidelines, enforce ethical standards, and ensure clarity in its application.
Author(s) Name: Satirtha Basak (Faculty of Law, Delhi University)
