Centre Number: PK243

Candidate Number: 0062

Candidate Name: Amna Majid

Topic 4: Digital World

Research Question: Is the use of A.I in education mostly justified in the 21st century?

Word Count: 1999

INTRODUCTION:

Artificial Intelligence (A.I) is computer software made to replicate human intelligence. A.I

based chatbots generate unique answers using a set database, where they find relevant

information to give to the user. Students can ask chatbots a variety of questions, such as, "What

is the meaning of life?" or "What year did Arizona become a state?" Being a conversational

chatbot, users can ask for further information or precise answers.

Regarding education, students use chatbots (e.g chatGPT) to summarise text, explain concepts,

write essays, etc. According to the June 2023 ACT survey (from 4,006 students in grades

10 to 12), 60% of students admit to the use of AI tools. This raises the question of where

the line should be drawn between AI assistance & AI misuse for students. This report will

explore the implications of AI both positive & negative from an educational perspective.

FOR:

This perspective believes that using chatbots should be allowed in education as firstly they

provide individual attention & secondly tackling the shortage of teachers.

Personalised attention is extremely beneficial for students, as it enables them to learn at

their own pace. Struggling students may require more attention, or students may be shy while

asking questions. By using A.I as an aid to learning concepts at their own pace, each

student would be able to use their full potential, which couldn't be possible for one teacher to

carry out. As a consequence, AI for educational support is growing in popularity, with 90%

of students who study with ChatGPT finding it better than a tutor. Each child has a different

way of processing information so AI can cater to that better than teachers because it generates

responses exactly how the user wants, e.g. learning business studies by referencing pop-

culture to help the student relate to the concept potentially increasing their understanding.

Therefore, by adjusting itself to the requirements of the user A.I has proved to be a great aid to

learning for students who prefer individual attention.

The source Intelligent served 3,017 high-school & college students (ages 16-24) along

with 3,234 parents of younger students in May 2023 asking about their experience of ChatGPT

as a tutor. This is a free, privately-supported website that’s editorially independent. That keeps

the authors details anonymous to encourage them to give candid feedback of their surveys and

write their findings without fear of any repercussions hence we can safely assume that this

survey of 6000 users fairly depicts the experience of the population it represents and can be

relied upon

The source URBE university ranked top 500 US isn’t as credible as the author Abel

Guerra is a marketing specialist rather than a professor, therefore his research depth is

questionable with his lack of expertise in IT/A.I research, with no proof or cross referencing in

the articles his value judgements can’t be taken as factual information.

Secondly, using AI in education can help tackle the shortage of teachers. In Pakistan the current

student-to-teacher ratio is 1:41 instead of an ideal 1:30 because 43% government teachers lack

basic training. This shortage is caused by lack of economic incentive, extensive workload

and the lack of professional staff. Thus, to tackle the shortage of qualified educators AI can

be used to create animated instructors powered by dynamic AI that can’t only deliver the

lecture to the students but also answer their basic questions & forward their advanced questions

to the area expert who can then either send in the answer to the student using the same chatbot

or can schedule a personalised e-session with him if required. Sending an A.I to schools can be

more cost effective as multiple tasks can be performed by one A.I bot which would require

many staff members for example, grading, scheduling, data management, & tutoring, so AI can

perform all of these tasks without getting tired unlike a human teacher allowing it to run

efficiently for a longer time. Consequently educational institutions can save costs by cutting

down resources required to operate efficiently, thereby increasing cost-effectiveness.

Therefore, AI powered digital teachers can embody teachers in places lacking resources for

educators, in turn creating availability of education for more places while keeping running

costs low.

The source published by ClassPoint, a company using A.I to create a product helping

educators, isn’t as reliable because they may have a vested interest to show that A.I is beneficial

in education to attract more customers to buy their software. The author Dianne Adlawan is a

Marketing Executive at ClassPoint isn’t reliable as it isn’t in her direct interest to show accurate

information, rather being a Marketing Executive it’s more likely for her to only show the

positive of AI inorder to attract customers.

The first argument of AI providing individualised learning is more important as it has proven

impacts for students, which is supported by the survey conducted by Intelligent. It is also more

realistic, as students have been using AI as private tutors. The second argument can’t be backed

by real-life examples as it’s mostly theoretical, however this argument does solve a major issue

present in the education sector, but it may not be feasible. Hence, the first argument is better as

it provides visible evidence to prove how A.I has benefited students by being used as private

tutor, whilst the second argument isn’t grounded in reality.

AGAINST:

This perspective believes that using chatbots shouldn't be encouraged in education because

they hinder the learning ability of students & privacy concerns while using AI.

The first argument is centred on AI hindering the learning ability of students. When students

students are given a tool capable of generating answers then it removes the premise of learning. Dean of Harvard’s Graduate School of Education Martin West believes that AI can undermine learning as students are no longer required to analyse & deduct answers of their own,❑[ 14 ] this is caused by the way chatbots generate answers; as they spit-out exact answers without requiring the student to conclude for themself❑[ 15 ], even if they believe they use chatbots to enhance or supplant their analytical thought, they are creating a facade for themselves or their teachers whereby they aren’t learning the material. Consequently, by using AI students bypass the learning process and directly get the answer, this harms their learning ability long term because they don’t develop the deductive skills required to analyse & understand concepts, instead they get used to getting straightforward answers which may give them better grades short-term, but for long-term they are hindering the learning ability of students.

To prevent students from copy-pasting answers for questions which require going from A to B, teachers should instead give students questions which challenge their intuition & personal reasoning. The questions should be aimed at testing the core concept of students rather than the simple implementation of the concepts which requires the student to go from A to B. For example, GP is a subject which tests a person's ability to reason, with questions requiring students to use their personal deductive reasoning which can’t be replaced by AI especially with the advancement of AI detection software such as GLTR, created by Harvard & MIT, ensures the authenticity of student work while discouraging usage of AI-generated content.

“GLTR AI Detection software made by Harvard & MIT has 72% without any prior training” ❑[ 16 ]. This research by Harward Independent from October 2023, 6 months after the launch of ChatGPT 4 gives sufficient time for researchers to conduct reasonable analysis on it, which is important due to the nature of my topic being very time-sensitive as constant changes are being made.

Moreover, chatbots raise privacy concerns for students in case data is leaked❑[ 17 ] , as students' personal data storage has increased significantly. According to Global cybersecurity & digital privacy company Kaspersky, the number of overall cyber threats in Pakistan has increased by 17% in 2023 as compared to 2022. ❑[ 18 ] , inevitably posing privacy concerns. This is because AI chatbots store the pattern of questions asked by students, which can be used to predict future behaviour❑[ 19 ] & collect data without their knowledge❑[ 20 ] going against the privacy rights of users. User data is collected through basic tasks such as generating documents, asking personal questions which could link someone to a place etc. According to the Cisco 2023 Consumer Privacy Survey❑[ 21 ] 39% users have entered work information, and over 25% have entered personal information, account numbers and even health information. So, especially students are at risk of unintentionally entering personal information which could be used by advertisers, commit identity theft and fraud ❑[ 22 ] . According to the same survey 88% users are concerned that the data they entered could be shared to third parties. UNICEF states that there is widespread belief data collection of minors should be protected at a higher level, but currently there is no agreed global data governance framework addressing concerns related to children’s data collection.❑[ 23 ] And the EdTech market is predicted to reach $404 billion by 2025 raising data security concerns as the economic incentive to harvest student data increases due to the increase in the market.❑[ 24 ] ; it’s clear that all data entered into a generative AI can be used by third parties, with UNICEF even claiming that currently there are no laws to prevent the data collection of minors. Hence, using generative AI imposes privacy concerns for students as their entries can be used by advertisers, commit identity theft, fraudulent activities or even track a person.

This source❑[ 18 ] by Tribune, is reliable as being a popular news outlet their publishing standard needs to be high, if they presented misleading information they can be openly criticised, therefore it’s in their interest to publish accurate information to ensure their credibility.

The first argument around students' over-reliability on chatbots is most important, as it addresses an issue that directly impacts the learning ability of students who tend to use these chatbots as an “easy way out”, but it permanently affects their research & analytical skills. It provides long-term consequences which if not addressed could affect students' grades & learning ability for their whole life because students learn & hone these basic skills in their younger years which are spent in schools. While data privacy & security are also important, they hold less value for students, as although for businesses leaking of data can be detrimental, for students it doesn’t have the same level of consequences. This is because students will generally use these A.I bots for educational purposes rather than personal use, making their data collection less harmful.

CONCLUSION:

As a student I found chatbots to be useful, but with prolonged use I realised the flaws & how it degraded my work as I was bound by the limits of chatbots and stopped finding fresh perspectives or thinking for myself which resulted in a poorer understanding of concepts.

4

According to this report, both sides have a similar scale of impact on students around the world but mainly in developed nations due to technological development. For arguments

arguments provide

significant benefits such as providing individualised attention & tackling the shortage of

teachers, but the second argument depends whether governments feel the need to use these

digital teachers as although they are cost effective the initial setting up costs & potential user

experience can hinder its overall results. The argument isn't backed by concrete evidence due

to their theoretical nature so it makes them less believable. On the other hand, opposing

arguments of student privacy & ethical concerns have more reliable sources and their effects

can be seen in the real world with students becoming more reliable on quick information

provided by these chatbots. Therefore, to fairly conclude more data is required to check the

long term effects of A.I in student behaviour/skills along with studies to prove whether or not

it’s feasible for countries to start using A.I powered instructors, future laws being created to

ensure data security for students.