October 17, 2025 • 25 mins
As part of our ongoing collaboration with Central Square Foundation, we are excited to bring to you the fourth episode of our five part series where we talk about the evolving landscape of AI in Education.
The National Education Policy 2020 marks a bold shift in how we think about technology in learning. It envisions a future where students build not just digital literacy, but also computational thinking and AI fluency — and where teachers are empowered with the tools, training, and support to integrate AI into their curriculums meaningfully and responsibly.
To understand how this is being implemented, we'll be joined by Gouri Gupta, Sr. Project Director of EdTech who leads CSF’s work in EdTech and AI and Professor Balaraman Ravindran, Head, Wadhwani School of Data Science & AI (WSAI), IIT Madras who is one of India’s top AI researchers and has helped shape India’s AI policy framework and currently advises the Reserve Bank of India on the uses of AI in finance.
Hosted and produced by Niharika Nanda
Edited and mixed by Suresh Pawar
Links to the previous episodes of our series with CSF:
Episode 1
Episode 2
Episode 3
The National Education Policy 2020 marks a bold shift in how we think about technology in learning. It envisions a future where students build not just digital literacy, but also computational thinking and AI fluency — and where teachers are empowered with the tools, training, and support to integrate AI into their curriculums meaningfully and responsibly.
To understand how this is being implemented, we'll be joined by Gouri Gupta, Sr. Project Director of EdTech who leads CSF’s work in EdTech and AI and Professor Balaraman Ravindran, Head, Wadhwani School of Data Science & AI (WSAI), IIT Madras who is one of India’s top AI researchers and has helped shape India’s AI policy framework and currently advises the Reserve Bank of India on the uses of AI in finance.
Hosted and produced by Niharika Nanda
Edited and mixed by Suresh Pawar
Links to the previous episodes of our series with CSF:
Episode 1
Episode 2
Episode 3
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:01):
Hi, I'm Naharikan Nanda and you are listening to Express podcasts.
The National Education Policy twenty twenty marks a bold shift
in how we think about technology in learning. It envisions
a future where students build not just digital literacy, but
also computational thinking and AI fluency, and where teachers are
(00:22):
impowered with the tools, training, and support to integrate artificial
intelligence into their curriculums meaningfully and responsibily. In the fourth
episode of our series, in association with Central Square Foundation,
we are going to discuss the evolving landscape of AI
in education. Recent advancements in AI have unlocked capabilities that
(00:44):
hold immense potential to catalyze innovation in education. There is
emerging evidence on AI in education for solutions that personalize
learning to the child's level or support teachers with pedagogical
resources from intelligent assessments that offer real time feedback to
enhanced lesson planning. The potential of AI to transform education
(01:07):
is immense. The lasting impact of AI is already being
felt in classrooms worldwide, with AI poys to reshape the
way we learn, teach, and innovate. In this episode, we
will try to understand prominent use cases of how AI
is already transforming teaching and learning in India, debunk common
myths around AI in classrooms and identify pathways to make
(01:30):
AI adoption inclusive and scalable in government schools. Joining us
today is Gory Gupta, Senior Project Director of ETTECH who
leads CSF's work in ETTECH and artificial intelligence, and Professor
Balaraman Rawindrin head Wadhwani School of Data Science and AI
itIt Madras. He is one of India's top AI researchers
(01:53):
and has helped shape India's AI policy framework and currently
advisors the Reserve Bank of India on the US users
of AI in finance. So, Gory, let's start with you.
What are the most promising ways in which AI can
be leveraged to reimagine education and to strengthen or enhance
the teaching learning experience of children in classrooms? And how
(02:15):
are you at CSF thinking about this.
Speaker 2 (02:18):
That's a great question in America. Thank you. I think
we are at a very interesting inflection point of discovery
because there are multiple experiments which are happening all over
the world to figure out how AI can be leveraged
to reimagine education. But I think the one unanimous opinion
(02:38):
that is coming out is that it is not going
to replace teachers, but it can help amplify their impact.
And I do believe that recent advancements AI have unlocked
capabilities that hold tremendous potential to catalyze innovation. Right, And
let me give you a few examples. So, one of
the most impactful interventions in education is personal due today,
(03:03):
and there have been multiple studies that demonstrate that it is,
you know, one of the most evidence backed interventions in education.
It was never possible to implement it at scale earlier, right,
But now with advancement in generative AI, we can think
about personalizing education for children at scale, right, and hence,
(03:26):
you know, it becomes a very powerful enabler to provide
this personalized, actionable feedback, providing analysis of recordings of classroom practices,
et cetera. The other place where I think AI can
be extremely powerful is in providing support to teachers through
personalized coaching. But with generative AI, there is a possibility
(03:48):
for her teacher to record a classroom and to get
feedback that is only meant for herself, right, for her
to improve. So these are two very impactful interventions that
we see. The third one is essentially around being able
to assess where a child is at a very high
frequency and being able to then tailor instruction bases the
(04:09):
performance of the class right, and this is called AI
powered Assessment Informed Instruction, another very impactful u SCCE that
can really help improve learning in the classroom. Having said that,
I do want to call out that given that this
technology is so new, you know, we still need to
figure out the evidence for what is actually working, how
is it working, and how it can it be delivered
(04:31):
at scale. And it is in that context that you know,
responding to your question of how CESF is thinking about it,
we are saying two things right. One is that we
already had a lot of interesting at tech which was
helping improve student learning outcomes or helping the teacher transact
better in the classroom. There are ways in which generative
AI can actually power these etech solutions and make them
(04:54):
more personalized, allow them to scale faster in terms of
their ability to generate content. So that's the one use
case that we see of a generative AI, which is
to strengthen existing at tech solutions. And the second one
that we see is for it to be able to
actually create these new teaching learning experiences that I was describing.
But we also realize that this is such a new
(05:16):
and emergent technology, right that we can in no way
imagine all the possibilities with it. So we at CSF
are thinking of setting up at tech Innovation Hub, which
will you know, almost gather ideas from you know, young professionals, students,
et cetera on what is the possibility to build educational
(05:38):
interventions with AI? And you know, we hope to be
able to have another conversation in Aharika next year where
we have many, many more interesting use cases that we
can talk about in the context of education and GENI.
Speaker 1 (05:53):
Right, and as you said, the possibilities are truly endless.
And Professor Ravendren now coming to you. As we introduce
AI into classrooms, two common concerns often come up, the
first one being the belief that AI in education is
just another passing tech fad that promises more than it delivers,
(06:14):
and the second one being the fear that AI might
produce children's creativity if students start relying on AI to
write stories, solve problems, or generate ideas.
Speaker 2 (06:24):
What would you say to educational leaders.
Speaker 1 (06:26):
Who worry that AI in classrooms is just a latest
tech trend, and how do you see AI being used
in ways that enhance creativity and critical thinking rather than
replacing it, especially in the context of school education.
Speaker 3 (06:41):
So let me start with the first part of your question.
So we're talking about whether the AI is a passing
fad or not. In fact, I would go to the
extreme and say that education has been lowered is probably
over so just the usual classroom instructions and looking at
students working on assignments and learning through interactions in the
(07:04):
class and things like that, that's going to be soon
a thing of the past. Because what AI, the current
generation of AI tools have done, is given access to
enormous volumes of information. And the key point here, I mean,
of course Google gave you access to enormous volumes of information,
But what has AI done is does it in a
(07:24):
way that is so easy for you to interact with. Right,
you can have a conversation with your information, which was
not something that was possible earlier. And this way of
accessing and processing information is going it's very powerful. Right,
So this communication challenge, right is something that the current
generation of AI tools have overcome and the students are
going to use it, whether question is are we going
(07:46):
to guide them properly to use it or are we
just going to you know, become like a wild wild
West where they just you know, get exposed to all
kinds of information and all kinds of wrong ways of
using these tools.
Speaker 4 (07:56):
As Guri was pointing out earlier in her response.
Speaker 3 (07:58):
Right, there's so many different in which these tools are
helping both the teachers and the students learn. And it's
no way a passive fact. Even at IITY, they're starting
to rethink about how we give assignments, right, I mean,
it just doesn't make sense for me to give assignments
to students. Look that they can just go and you know,
start interacting with one of the char GPT models. In fact,
I even took some of the open ended questions I
(08:19):
used to give in the past and ran it past
these AI models and they were able to give very
reasonable answers. I mean there's an open ended questions, then
they have no right or wrong answers. Right, it's basically
to see how you reason through these issues. So I
had to stop giving assignments in class. So I mean
I had to make them sit in the class and
solve problems without access to an AI moal Right, So
that's I mean, we have to start to rethinking, you know,
how we do education.
Speaker 4 (08:40):
It's not a passing fact.
Speaker 3 (08:41):
It's going to fundamentally alter how classrooms would be structured going.
Speaker 4 (08:45):
Forward in the future.
Speaker 3 (08:46):
And the second question you asked us about creativity, right,
So there are different ways of looking at it. So
if you think about it, you know when people started
using calculators extensively, right, So, I mean there is always
this question that you're going to lose your fundamental human
I think you can't do arithmetic anymore and things like that,
(09:06):
and arithmetic has become a challenge. I mean, most kids
nowadays can't add numbers in their head, right, So people
of my generation can certainly do non trivial, you know, arithmetic.
So one way of passing this is to say that, hey, look,
it's not that you're going to become less creative. It's
just that you're going to be able to use AI
for you know, a little bit of the mundane things
(09:26):
and then think at a higher level. Right, So we
have to train people how to use these like just
like you use calculators, and then think about doing higher
mathematics like that, you should think about how would you
use AI and then you know, elevate your thought process,
elevate your creative process to a level which was not
possible earlier. You know, in lower grades you cannot use
a calculator. You have to do pen and paper arithmetic.
(09:48):
That has to be you know, a more gradual way
of introducing children to problems solving us seeing AI tools.
But first they have to have some training in the
Faculty of Political Thinking before they switched to starting to
use AI to solve problem.
Speaker 2 (10:03):
And Professor Ravendo you can just add, right, we keep
doing these dipstick surveys as part of our work in
CSF to figure out, you know, how students and teachers
are using AI, and we see significant shifts happening in
a three month period. Right. So three months ago, for example,
you know, we did a survey and we found that
there was you know, usage of AI, but limited awareness
(10:25):
around it. Three months later, we find more rampant usage,
more awareness. So this is a technology that you know,
is getting embraced at much faster rate than what we
have ever seen before and hence, you know it almost
what you're saying is so important that you know, the
education system has to be one step ahead in figuring
out how to deal with it and what is a
(10:47):
guidance that needs to be provided?
Speaker 1 (10:49):
Absolutely, and I feel the key word here, as you
mentioned is elevate to use AI to elevate the skills
that can be developed through education and Professor Ravenrin, we
are beginning to see real promise in how AI could
help us tackle some of education's toughest challenges. But when
we bring this conversation into the context of India's public
(11:12):
school system, there is often a perception that building or
using AI, especially gen AI, is simply too expensive or
out of reach for government classrooms. So what would you
say about this particular myth that AI is only for
affluent users and that most tools are premium priced and
out of reach for most of India.
Speaker 3 (11:34):
I think it's misconception that accessing AI is out of
reach of most of India.
Speaker 4 (11:40):
So that is AI, and that is AI.
Speaker 3 (11:42):
I mean, it's not like every student learner needs to
have access to the AI that can crack the math
Olympia problems, right, So we are really looking at you know,
AI that helps them to learn and makes the whole
the learning process.
Speaker 4 (11:56):
Much smoother, so in that sense. Right.
Speaker 3 (11:58):
So the twenty twenty four Annual State of Education Survey
Report right said that nearly eighty two percent of children
in rural India between ages fourteen to sixteen already use
smartphones and have access to some form of AA or
the other right, whether it is through chat pots, whether
you're talking to your phone.
Speaker 4 (12:15):
Right.
Speaker 3 (12:15):
So whenever the people are using a voice interface to
search for something, right, they.
Speaker 4 (12:20):
Are using AI, right, whether they know it or not.
Speaker 3 (12:22):
So there's a lot of ways in which they are
already interacting with AI tools and so it's not the access, right,
it's all they are using it to learn, right, That
becomes more of a challenge. So when we start thinking about,
you know, a school giving more systematic access to AI system,
So we did some back of the you know, enveloped
computation in some sense right, so both at CSF and
(12:42):
c I by the Center for Responsible AA that I had,
So we did some work on this and then we
figured out that if you're using a medium level commercial tool, right,
not necessarily the most expensive of these tools. If using
A for example, you don't need access to the latest
greatest Dept five, Right, you could use a version that is,
you know N minus one or N minus two version
(13:03):
that is out there, and then having students work with this.
Speaker 4 (13:06):
It can be roughly as.
Speaker 3 (13:08):
Slow as about one hundred to one hundred ten repiece
per student per year if you are thinking about a
scale of about ten million students, and this can give
the students an access of about forty five minutes no
interaction time with an AI every week and with about
ten minutes of chatboard interactions.
Speaker 4 (13:23):
Right.
Speaker 3 (13:24):
So, even even using this high level AI, I mean,
if you kind of structure your curriculum, structure the assignments
in such a way, this is good to give the
students a significant exposure to how to use AI to learn.
But then the true inclusion, true impact would come when
we start building our own sovereign models, right and then
a lot of effort is going towards that in India
(13:45):
right now. And so if you start looking at our
own sovereign models, you can actually bring the cost down
to you know, roughly twenty five to thirty piece per
student per.
Speaker 4 (13:54):
Year and provide about one hour.
Speaker 3 (13:57):
Of AII US stage per day, assuming that we are
able to invest the significant amount of prend right, I mean,
it will take a few crores maybe fourteen fifteen crores
of investment upfront, but you know this will get amorphlazed
over several years. Hey, let's say reasonably we can take
a five year window on which you know you would
like to provide these services based on the models that
we build, so in which case you can give you know,
(14:17):
daily access to the students, non trivial daily access to
students to use AI and this can certainly have a
much much broader impact, right.
Speaker 1 (14:25):
Right, and professor, what are the efforts being undertaken by
the government and academia to make AI education truly accessible
and scalable across the public system?
Speaker 3 (14:36):
So there are multiple initiatives that are being taken. So
the biggest challenge is in the Indian ecosystem, Right is
the access to compute, access to data and access to
sovereign models. Right, So, sovereign models being ones that you know,
there are good in Indian languages that are going to
be trained on Indian data and run on Indian systems
and Indian hardware and so and access is to India first, right,
(14:58):
So these are what we call sovereign mode in some sense.
And so access to sovereign models, access to actual training
data and access to computer resources have always been a challenge,
so the government is addressing each one of these directly, right.
So there's a lot of effort that has been now
done to provide subsidies access to GPUs for people for
trying to build the sovereign models. And then there's also
several funding initiatives by both the Ministry of It and
(15:21):
as well as the Ministry of Education towards you know,
building models or tailor to specific applications. The Ministry of
Education is now funding AI Center of Excellence where the
goal would be to build these models to help address
a FLN, particularly targeted towards government schools in rural areas
where the literacy component has to be addressed in the
(15:42):
local language, so whatever is the language that's spoken at home, right,
So that's the goal, and they are basically trying to
push both academic research and in the industry option in
that direct So that's something done by the Ministry of Education.
And likewise, there's a lot of effort in building data
sets right, so Indian specific data sets and for building
these foundation models. So quite a bit of effort is
being put into that space. And one nice thing about
(16:03):
the latest round of funding that the government has been doing.
I mean this is non trivial amount, right, I'm talking
about several hundred crows and I mean a couple of
thousand crows from the Ministry of It and so on and
so forth. The goal has always been to produce technology
that is deployable at scale across the country, not just
in well connected cities but also in rural areas.
Speaker 4 (16:24):
So I think this focus.
Speaker 3 (16:25):
That the government is pushing everyone to look at right
will have a much much broader impact. In fact, we
participated in this call for this educatione AIR for Education
CEOE and the questions that you know we had to address, well,
we put together our proposal. We're all about you know, scaling.
So hey, everybody knows that air can be used to
solve problem. You don't want to hear that. How will
you scale it to the nations level?
Speaker 1 (16:46):
Right?
Speaker 3 (16:46):
How will you scale it to rural areas where the
connectivity is going to be challenged and things like that.
So I think the government is on the right path
encouraging both industry and the academia to think about solving
problems for India, and a lot of the academic rests
particularly you know AI for Barath here and then I
bombay what efforts that they're doing. I think all of
this are basically focused on other thing, this Indian language
(17:09):
fluency in these models, right, So I think that's a
big push that's coming from the academic.
Speaker 1 (17:14):
Side, right And Gory now coming back to you, why
is equipping students, parents and teachers with AI literacy a
critical foundation for driving responsible and large scale AI adoption
in education? What is CSF doing in this regard?
Speaker 2 (17:32):
Yeah, So, Niharika, you just heard Professor of Indren talk
about the huge thrust and investment that is going into
ensuring that AI powered tools become available for education. Right,
and in order for teachers and students to be able
to responsibly adopt this right, AI literacy in our minds
becomes critical. Right, So, birth let me even start by
(17:55):
defining what AI literacy is, right, So, in our minds,
AI literacy refers to the knowledge and practice that help students, teachers,
and parents understand what AI is, how it impacts our
increasingly digital lives, and how we need to use it meaningful,
safely and responsibility. And I wanted to give this definition
(18:15):
in order to almost differentiate a little bit between AI
literacy and AI skills, right, which is about saying that, look,
I have the technical skills to build and create AI
versus I have the knowledge and the understanding to responsibly
use A And we believe that every educator, every student,
(18:36):
every parent in India needs to be AI literate because
AI is proliferating their lives in manners which is hard
to understand. So in this context, in collaboration with the
one one School of Data Science and AI at iit
Matras and supported by Google dot Org, CSF actually launched
an initiative called AI Someone. And there were multiple paths
(18:59):
to A someone, right. So there was one part to
say that, look, we need to build a curriculum or
government school students and teachers that is contextual for India, right,
that is contextual to the starting point of students and teachers, right.
And that was the first thing that we did, which
was a curriculum built and that curriculum is available on
(19:19):
the aisumm at the website. The second part was to
build learning stax around this, right, because there's not as
much content for EI literacy in the Indian languages suited
to the Indian context as we would want it to
be right. So they're learning stacks that are built for
both teachers and students that can be used in online
modes but can also be used to deliver this instruction
(19:42):
in the classroom. And then we wanted to actually scale
this initiative across India and hence we partnered with organizations
that are working currently in middle school so grade six
to nine across you know, multiple states in India to
deliver this curriculum and content and to five million students
and teachers in the first phase and hoping to really
(20:05):
scale this up across Inda. So we believe that this
combination of building out a curriculum, the contents stack associated
with the curriculum, and ensuring that it gets adopted at
scale will at least, you know, be the first step
in introducing AI in literacy within the education system in India.
Speaker 4 (20:22):
Can I and one point to that.
Speaker 3 (20:24):
So CPAC is very keen on introducing AI into the
curriculum starting from the third grade. So in fact, we
have been having conversations with them and trying to figure
out what it means and want to start introducing kids
in the third grade. And so we have been proposing
a very gradual introduction to AI from a very early
stage in their learning, so that they kind of you know,
(20:44):
because even at the third grade level, the students are
starting to interact with AI, so they really need to
know at least the best practices how to stay safe
and you're interacting with air models, and you know, things
like how not to have any transparents of emotional attachment AI,
and so those are the very fundamental stuff, even not
the early stage that kids should start understanding. So it's
going to be Societal infractions are already changing because of
(21:06):
AI model, right, people are isolating themselves more and more,
and especially teenagers and interacting their Also, I think literacy,
both at early ages for the children as well as
the parents and teachers are becoming more and more.
Speaker 2 (21:18):
Right and Gody.
Speaker 1 (21:20):
A common myth we encounter is that AI literacy isn't
a priority for children from low income communities, that we
should first fix the basics before introducing something as advanced
as artificial intelligence. But this assumes AI is a distant
or optional concern. In reality, children from all backgrounds are
(21:41):
already engaging with AI. As Professor Revendron also mentioned, whether
it is through voice assistance or recommendation algorithms or automated
chat pots without necessarily understanding how these systems work or
impact their choices. So my question is, how do we
bust this myth and build at wrong case for AI
literacy as an equity issue, not just for future skills agenda,
(22:06):
And what should we keep in mind to make AI
literacy truly meaningful and accessible for learners from underserved communities,
you know, just.
Speaker 2 (22:14):
Following up from Professor ND's point, also is the fact
that you know, with AI literacy, we are not talking
about children holding neural networks, right, we are talking about
helping them recognize when they're interacting with AI, how are
these algos shape what they see and hear, how to
question them and use these systems responsibly andthetically right in
(22:37):
that sense, I mean, in my mind it's as fundamental
as you know, any form of digital literacy that we
need today, right, and for low income learners, in my mind,
it's it's almost like a justice issue because the truth
is that your children in these communities, and we've been
saying it over and over again, right, they're all ready
users of AI, but they are really the shapers of
(22:59):
this right and global frameworks whether it's the UNSCO SKATE
twelve AI Competency Model or the OEC the AI Literacy guidelines.
They all stress on the fact that equitable AI education
must begin early, right, And also, going back to Professor
of Inuran's point, are the fact that CBA is wanting
to induce this early. But so you know, we need
(23:20):
to ensure that, you know, children going to government schools
affordable private schools are able to understand and influence this technology, right,
rather than simply being the passive consumers who get acted
upon by this technology. And hence it becomes you know,
extremely important for us to ensure that AI literacy is
(23:42):
something that is introduced for all children and educators and
their parents as early as possible, right. And the idea
is not to teach AI as a new subject, right,
but to almost nurture that consciousness about AI right, helping
these children see themselves as participants in this technology. Otherwise
(24:02):
we will further be wiping the gap right between children
and teachers in school systems that we'll be using this
technology right to do what Professor Vendron was describing as
do more right, and those who are just unaware and
not being able to use it well, and hence you know,
just to say what you were saying, Ahika right, in
my mind, it becomes an equity agenda because it's not
(24:25):
just about access, it's about you know, the agency to
be able to use it.
Speaker 3 (24:29):
Well.
Speaker 1 (24:33):
You were listening to Gory Gutta and Bala Raman Ravendren
in the fourth episode of our five part series in
association with CSF.
Speaker 2 (24:41):
This episode was.
Speaker 1 (24:42):
Edited and mixed by Seshbawar and produced by me and
Hari Kananda. If you want to listen to the previous
episodes of this series, the links to the same are
provided in the description. If you like the show, then
do subscribe to us wherever you get your podcasts. You
can also recommend the show to some you think we'll
like it, share it with a friend or someone in
(25:03):
your family. It's the best way for people to get
to know about us. You can also tweet us at
Express Podcasts or write to us at podcasts at innie
Express dot com.
Hi, I'm Naharikan Nanda and you are listening to Express podcasts.
The National Education Policy twenty twenty marks a bold shift
in how we think about technology in learning. It envisions
a future where students build not just digital literacy, but
also computational thinking and AI fluency, and where teachers are
(00:22):
impowered with the tools, training, and support to integrate artificial
intelligence into their curriculums meaningfully and responsibily. In the fourth
episode of our series, in association with Central Square Foundation,
we are going to discuss the evolving landscape of AI
in education. Recent advancements in AI have unlocked capabilities that
(00:44):
hold immense potential to catalyze innovation in education. There is
emerging evidence on AI in education for solutions that personalize
learning to the child's level or support teachers with pedagogical
resources from intelligent assessments that offer real time feedback to
enhanced lesson planning. The potential of AI to transform education
(01:07):
is immense. The lasting impact of AI is already being
felt in classrooms worldwide, with AI poys to reshape the
way we learn, teach, and innovate. In this episode, we
will try to understand prominent use cases of how AI
is already transforming teaching and learning in India, debunk common
myths around AI in classrooms and identify pathways to make
(01:30):
AI adoption inclusive and scalable in government schools. Joining us
today is Gory Gupta, Senior Project Director of ETTECH who
leads CSF's work in ETTECH and artificial intelligence, and Professor
Balaraman Rawindrin head Wadhwani School of Data Science and AI
itIt Madras. He is one of India's top AI researchers
(01:53):
and has helped shape India's AI policy framework and currently
advisors the Reserve Bank of India on the US users
of AI in finance. So, Gory, let's start with you.
What are the most promising ways in which AI can
be leveraged to reimagine education and to strengthen or enhance
the teaching learning experience of children in classrooms? And how
(02:15):
are you at CSF thinking about this.
Speaker 2 (02:18):
That's a great question in America. Thank you. I think
we are at a very interesting inflection point of discovery
because there are multiple experiments which are happening all over
the world to figure out how AI can be leveraged
to reimagine education. But I think the one unanimous opinion
(02:38):
that is coming out is that it is not going
to replace teachers, but it can help amplify their impact.
And I do believe that recent advancements AI have unlocked
capabilities that hold tremendous potential to catalyze innovation. Right, And
let me give you a few examples. So, one of
the most impactful interventions in education is personal due today,
(03:03):
and there have been multiple studies that demonstrate that it is,
you know, one of the most evidence backed interventions in education.
It was never possible to implement it at scale earlier, right,
But now with advancement in generative AI, we can think
about personalizing education for children at scale, right, and hence,
(03:26):
you know, it becomes a very powerful enabler to provide
this personalized, actionable feedback, providing analysis of recordings of classroom practices,
et cetera. The other place where I think AI can
be extremely powerful is in providing support to teachers through
personalized coaching. But with generative AI, there is a possibility
(03:48):
for her teacher to record a classroom and to get
feedback that is only meant for herself, right, for her
to improve. So these are two very impactful interventions that
we see. The third one is essentially around being able
to assess where a child is at a very high
frequency and being able to then tailor instruction bases the
(04:09):
performance of the class right, and this is called AI
powered Assessment Informed Instruction, another very impactful u SCCE that
can really help improve learning in the classroom. Having said that,
I do want to call out that given that this
technology is so new, you know, we still need to
figure out the evidence for what is actually working, how
is it working, and how it can it be delivered
(04:31):
at scale. And it is in that context that you know,
responding to your question of how CESF is thinking about it,
we are saying two things right. One is that we
already had a lot of interesting at tech which was
helping improve student learning outcomes or helping the teacher transact
better in the classroom. There are ways in which generative
AI can actually power these etech solutions and make them
(04:54):
more personalized, allow them to scale faster in terms of
their ability to generate content. So that's the one use
case that we see of a generative AI, which is
to strengthen existing at tech solutions. And the second one
that we see is for it to be able to
actually create these new teaching learning experiences that I was describing.
But we also realize that this is such a new
(05:16):
and emergent technology, right that we can in no way
imagine all the possibilities with it. So we at CSF
are thinking of setting up at tech Innovation Hub, which
will you know, almost gather ideas from you know, young professionals, students,
et cetera on what is the possibility to build educational
(05:38):
interventions with AI? And you know, we hope to be
able to have another conversation in Aharika next year where
we have many, many more interesting use cases that we
can talk about in the context of education and GENI.
Speaker 1 (05:53):
Right, and as you said, the possibilities are truly endless.
And Professor Ravendren now coming to you. As we introduce
AI into classrooms, two common concerns often come up, the
first one being the belief that AI in education is
just another passing tech fad that promises more than it delivers,
(06:14):
and the second one being the fear that AI might
produce children's creativity if students start relying on AI to
write stories, solve problems, or generate ideas.
Speaker 2 (06:24):
What would you say to educational leaders.
Speaker 1 (06:26):
Who worry that AI in classrooms is just a latest
tech trend, and how do you see AI being used
in ways that enhance creativity and critical thinking rather than
replacing it, especially in the context of school education.
Speaker 3 (06:41):
So let me start with the first part of your question.
So we're talking about whether the AI is a passing
fad or not. In fact, I would go to the
extreme and say that education has been lowered is probably
over so just the usual classroom instructions and looking at
students working on assignments and learning through interactions in the
(07:04):
class and things like that, that's going to be soon
a thing of the past. Because what AI, the current
generation of AI tools have done, is given access to
enormous volumes of information. And the key point here, I mean,
of course Google gave you access to enormous volumes of information,
But what has AI done is does it in a
(07:24):
way that is so easy for you to interact with. Right,
you can have a conversation with your information, which was
not something that was possible earlier. And this way of
accessing and processing information is going it's very powerful. Right,
So this communication challenge, right is something that the current
generation of AI tools have overcome and the students are
going to use it, whether question is are we going
(07:46):
to guide them properly to use it or are we
just going to you know, become like a wild wild
West where they just you know, get exposed to all
kinds of information and all kinds of wrong ways of
using these tools.
Speaker 4 (07:56):
As Guri was pointing out earlier in her response.
Speaker 3 (07:58):
Right, there's so many different in which these tools are
helping both the teachers and the students learn. And it's
no way a passive fact. Even at IITY, they're starting
to rethink about how we give assignments, right, I mean,
it just doesn't make sense for me to give assignments
to students. Look that they can just go and you know,
start interacting with one of the char GPT models. In fact,
I even took some of the open ended questions I
(08:19):
used to give in the past and ran it past
these AI models and they were able to give very
reasonable answers. I mean there's an open ended questions, then
they have no right or wrong answers. Right, it's basically
to see how you reason through these issues. So I
had to stop giving assignments in class. So I mean
I had to make them sit in the class and
solve problems without access to an AI moal Right, So
that's I mean, we have to start to rethinking, you know,
how we do education.
Speaker 4 (08:40):
It's not a passing fact.
Speaker 3 (08:41):
It's going to fundamentally alter how classrooms would be structured going.
Speaker 4 (08:45):
Forward in the future.
Speaker 3 (08:46):
And the second question you asked us about creativity, right,
So there are different ways of looking at it. So
if you think about it, you know when people started
using calculators extensively, right, So, I mean there is always
this question that you're going to lose your fundamental human
I think you can't do arithmetic anymore and things like that,
(09:06):
and arithmetic has become a challenge. I mean, most kids
nowadays can't add numbers in their head, right, So people
of my generation can certainly do non trivial, you know, arithmetic.
So one way of passing this is to say that, hey, look,
it's not that you're going to become less creative. It's
just that you're going to be able to use AI
for you know, a little bit of the mundane things
(09:26):
and then think at a higher level. Right, So we
have to train people how to use these like just
like you use calculators, and then think about doing higher
mathematics like that, you should think about how would you
use AI and then you know, elevate your thought process,
elevate your creative process to a level which was not
possible earlier. You know, in lower grades you cannot use
a calculator. You have to do pen and paper arithmetic.
(09:48):
That has to be you know, a more gradual way
of introducing children to problems solving us seeing AI tools.
But first they have to have some training in the
Faculty of Political Thinking before they switched to starting to
use AI to solve problem.
Speaker 2 (10:03):
And Professor Ravendo you can just add, right, we keep
doing these dipstick surveys as part of our work in
CSF to figure out, you know, how students and teachers
are using AI, and we see significant shifts happening in
a three month period. Right. So three months ago, for example,
you know, we did a survey and we found that
there was you know, usage of AI, but limited awareness
(10:25):
around it. Three months later, we find more rampant usage,
more awareness. So this is a technology that you know,
is getting embraced at much faster rate than what we
have ever seen before and hence, you know it almost
what you're saying is so important that you know, the
education system has to be one step ahead in figuring
out how to deal with it and what is a
(10:47):
guidance that needs to be provided?
Speaker 1 (10:49):
Absolutely, and I feel the key word here, as you
mentioned is elevate to use AI to elevate the skills
that can be developed through education and Professor Ravenrin, we
are beginning to see real promise in how AI could
help us tackle some of education's toughest challenges. But when
we bring this conversation into the context of India's public
(11:12):
school system, there is often a perception that building or
using AI, especially gen AI, is simply too expensive or
out of reach for government classrooms. So what would you
say about this particular myth that AI is only for
affluent users and that most tools are premium priced and
out of reach for most of India.
Speaker 3 (11:34):
I think it's misconception that accessing AI is out of
reach of most of India.
Speaker 4 (11:40):
So that is AI, and that is AI.
Speaker 3 (11:42):
I mean, it's not like every student learner needs to
have access to the AI that can crack the math
Olympia problems, right, So we are really looking at you know,
AI that helps them to learn and makes the whole
the learning process.
Speaker 4 (11:56):
Much smoother, so in that sense. Right.
Speaker 3 (11:58):
So the twenty twenty four Annual State of Education Survey
Report right said that nearly eighty two percent of children
in rural India between ages fourteen to sixteen already use
smartphones and have access to some form of AA or
the other right, whether it is through chat pots, whether
you're talking to your phone.
Speaker 4 (12:15):
Right.
Speaker 3 (12:15):
So whenever the people are using a voice interface to
search for something, right, they.
Speaker 4 (12:20):
Are using AI, right, whether they know it or not.
Speaker 3 (12:22):
So there's a lot of ways in which they are
already interacting with AI tools and so it's not the access, right,
it's all they are using it to learn, right, That
becomes more of a challenge. So when we start thinking about,
you know, a school giving more systematic access to AI system,
So we did some back of the you know, enveloped
computation in some sense right, so both at CSF and
(12:42):
c I by the Center for Responsible AA that I had,
So we did some work on this and then we
figured out that if you're using a medium level commercial tool, right,
not necessarily the most expensive of these tools. If using
A for example, you don't need access to the latest
greatest Dept five, Right, you could use a version that is,
you know N minus one or N minus two version
(13:03):
that is out there, and then having students work with this.
Speaker 4 (13:06):
It can be roughly as.
Speaker 3 (13:08):
Slow as about one hundred to one hundred ten repiece
per student per year if you are thinking about a
scale of about ten million students, and this can give
the students an access of about forty five minutes no
interaction time with an AI every week and with about
ten minutes of chatboard interactions.
Speaker 4 (13:23):
Right.
Speaker 3 (13:24):
So, even even using this high level AI, I mean,
if you kind of structure your curriculum, structure the assignments
in such a way, this is good to give the
students a significant exposure to how to use AI to learn.
But then the true inclusion, true impact would come when
we start building our own sovereign models, right and then
a lot of effort is going towards that in India
(13:45):
right now. And so if you start looking at our
own sovereign models, you can actually bring the cost down
to you know, roughly twenty five to thirty piece per
student per.
Speaker 4 (13:54):
Year and provide about one hour.
Speaker 3 (13:57):
Of AII US stage per day, assuming that we are
able to invest the significant amount of prend right, I mean,
it will take a few crores maybe fourteen fifteen crores
of investment upfront, but you know this will get amorphlazed
over several years. Hey, let's say reasonably we can take
a five year window on which you know you would
like to provide these services based on the models that
we build, so in which case you can give you know,
(14:17):
daily access to the students, non trivial daily access to
students to use AI and this can certainly have a
much much broader impact, right.
Speaker 1 (14:25):
Right, and professor, what are the efforts being undertaken by
the government and academia to make AI education truly accessible
and scalable across the public system?
Speaker 3 (14:36):
So there are multiple initiatives that are being taken. So
the biggest challenge is in the Indian ecosystem, Right is
the access to compute, access to data and access to
sovereign models. Right, So, sovereign models being ones that you know,
there are good in Indian languages that are going to
be trained on Indian data and run on Indian systems
and Indian hardware and so and access is to India first, right,
(14:58):
So these are what we call sovereign mode in some sense.
And so access to sovereign models, access to actual training
data and access to computer resources have always been a challenge,
so the government is addressing each one of these directly, right.
So there's a lot of effort that has been now
done to provide subsidies access to GPUs for people for
trying to build the sovereign models. And then there's also
several funding initiatives by both the Ministry of It and
(15:21):
as well as the Ministry of Education towards you know,
building models or tailor to specific applications. The Ministry of
Education is now funding AI Center of Excellence where the
goal would be to build these models to help address
a FLN, particularly targeted towards government schools in rural areas
where the literacy component has to be addressed in the
(15:42):
local language, so whatever is the language that's spoken at home, right,
So that's the goal, and they are basically trying to
push both academic research and in the industry option in
that direct So that's something done by the Ministry of Education.
And likewise, there's a lot of effort in building data
sets right, so Indian specific data sets and for building
these foundation models. So quite a bit of effort is
being put into that space. And one nice thing about
(16:03):
the latest round of funding that the government has been doing.
I mean this is non trivial amount, right, I'm talking
about several hundred crows and I mean a couple of
thousand crows from the Ministry of It and so on and
so forth. The goal has always been to produce technology
that is deployable at scale across the country, not just
in well connected cities but also in rural areas.
Speaker 4 (16:24):
So I think this focus.
Speaker 3 (16:25):
That the government is pushing everyone to look at right
will have a much much broader impact. In fact, we
participated in this call for this educatione AIR for Education
CEOE and the questions that you know we had to address, well,
we put together our proposal. We're all about you know, scaling.
So hey, everybody knows that air can be used to
solve problem. You don't want to hear that. How will
you scale it to the nations level?
Speaker 1 (16:46):
Right?
Speaker 3 (16:46):
How will you scale it to rural areas where the
connectivity is going to be challenged and things like that.
So I think the government is on the right path
encouraging both industry and the academia to think about solving
problems for India, and a lot of the academic rests
particularly you know AI for Barath here and then I
bombay what efforts that they're doing. I think all of
this are basically focused on other thing, this Indian language
(17:09):
fluency in these models, right, So I think that's a
big push that's coming from the academic.
Speaker 1 (17:14):
Side, right And Gory now coming back to you, why
is equipping students, parents and teachers with AI literacy a
critical foundation for driving responsible and large scale AI adoption
in education? What is CSF doing in this regard?
Speaker 2 (17:32):
Yeah, So, Niharika, you just heard Professor of Indren talk
about the huge thrust and investment that is going into
ensuring that AI powered tools become available for education. Right,
and in order for teachers and students to be able
to responsibly adopt this right, AI literacy in our minds
becomes critical. Right, So, birth let me even start by
(17:55):
defining what AI literacy is, right, So, in our minds,
AI literacy refers to the knowledge and practice that help students, teachers,
and parents understand what AI is, how it impacts our
increasingly digital lives, and how we need to use it meaningful,
safely and responsibility. And I wanted to give this definition
(18:15):
in order to almost differentiate a little bit between AI
literacy and AI skills, right, which is about saying that, look,
I have the technical skills to build and create AI
versus I have the knowledge and the understanding to responsibly
use A And we believe that every educator, every student,
(18:36):
every parent in India needs to be AI literate because
AI is proliferating their lives in manners which is hard
to understand. So in this context, in collaboration with the
one one School of Data Science and AI at iit
Matras and supported by Google dot Org, CSF actually launched
an initiative called AI Someone. And there were multiple paths
(18:59):
to A someone, right. So there was one part to
say that, look, we need to build a curriculum or
government school students and teachers that is contextual for India, right,
that is contextual to the starting point of students and teachers, right.
And that was the first thing that we did, which
was a curriculum built and that curriculum is available on
(19:19):
the aisumm at the website. The second part was to
build learning stax around this, right, because there's not as
much content for EI literacy in the Indian languages suited
to the Indian context as we would want it to
be right. So they're learning stacks that are built for
both teachers and students that can be used in online
modes but can also be used to deliver this instruction
(19:42):
in the classroom. And then we wanted to actually scale
this initiative across India and hence we partnered with organizations
that are working currently in middle school so grade six
to nine across you know, multiple states in India to
deliver this curriculum and content and to five million students
and teachers in the first phase and hoping to really
(20:05):
scale this up across Inda. So we believe that this
combination of building out a curriculum, the contents stack associated
with the curriculum, and ensuring that it gets adopted at
scale will at least, you know, be the first step
in introducing AI in literacy within the education system in India.
Speaker 4 (20:22):
Can I and one point to that.
Speaker 3 (20:24):
So CPAC is very keen on introducing AI into the
curriculum starting from the third grade. So in fact, we
have been having conversations with them and trying to figure
out what it means and want to start introducing kids
in the third grade. And so we have been proposing
a very gradual introduction to AI from a very early
stage in their learning, so that they kind of you know,
(20:44):
because even at the third grade level, the students are
starting to interact with AI, so they really need to
know at least the best practices how to stay safe
and you're interacting with air models, and you know, things
like how not to have any transparents of emotional attachment AI,
and so those are the very fundamental stuff, even not
the early stage that kids should start understanding. So it's
going to be Societal infractions are already changing because of
(21:06):
AI model, right, people are isolating themselves more and more,
and especially teenagers and interacting their Also, I think literacy,
both at early ages for the children as well as
the parents and teachers are becoming more and more.
Speaker 2 (21:18):
Right and Gody.
Speaker 1 (21:20):
A common myth we encounter is that AI literacy isn't
a priority for children from low income communities, that we
should first fix the basics before introducing something as advanced
as artificial intelligence. But this assumes AI is a distant
or optional concern. In reality, children from all backgrounds are
(21:41):
already engaging with AI. As Professor Revendron also mentioned, whether
it is through voice assistance or recommendation algorithms or automated
chat pots without necessarily understanding how these systems work or
impact their choices. So my question is, how do we
bust this myth and build at wrong case for AI
literacy as an equity issue, not just for future skills agenda,
(22:06):
And what should we keep in mind to make AI
literacy truly meaningful and accessible for learners from underserved communities,
you know, just.
Speaker 2 (22:14):
Following up from Professor ND's point, also is the fact
that you know, with AI literacy, we are not talking
about children holding neural networks, right, we are talking about
helping them recognize when they're interacting with AI, how are
these algos shape what they see and hear, how to
question them and use these systems responsibly andthetically right in
(22:37):
that sense, I mean, in my mind it's as fundamental
as you know, any form of digital literacy that we
need today, right, and for low income learners, in my mind,
it's it's almost like a justice issue because the truth
is that your children in these communities, and we've been
saying it over and over again, right, they're all ready
users of AI, but they are really the shapers of
(22:59):
this right and global frameworks whether it's the UNSCO SKATE
twelve AI Competency Model or the OEC the AI Literacy guidelines.
They all stress on the fact that equitable AI education
must begin early, right, And also, going back to Professor
of Inuran's point, are the fact that CBA is wanting
to induce this early. But so you know, we need
(23:20):
to ensure that, you know, children going to government schools
affordable private schools are able to understand and influence this technology, right,
rather than simply being the passive consumers who get acted
upon by this technology. And hence it becomes you know,
extremely important for us to ensure that AI literacy is
(23:42):
something that is introduced for all children and educators and
their parents as early as possible, right. And the idea
is not to teach AI as a new subject, right,
but to almost nurture that consciousness about AI right, helping
these children see themselves as participants in this technology. Otherwise
(24:02):
we will further be wiping the gap right between children
and teachers in school systems that we'll be using this
technology right to do what Professor Vendron was describing as
do more right, and those who are just unaware and
not being able to use it well, and hence you know,
just to say what you were saying, Ahika right, in
my mind, it becomes an equity agenda because it's not
(24:25):
just about access, it's about you know, the agency to
be able to use it.
Speaker 3 (24:29):
Well.
Speaker 1 (24:33):
You were listening to Gory Gutta and Bala Raman Ravendren
in the fourth episode of our five part series in
association with CSF.
Speaker 2 (24:41):
This episode was.
Speaker 1 (24:42):
Edited and mixed by Seshbawar and produced by me and
Hari Kananda. If you want to listen to the previous
episodes of this series, the links to the same are
provided in the description. If you like the show, then
do subscribe to us wherever you get your podcasts. You
can also recommend the show to some you think we'll
like it, share it with a friend or someone in
(25:03):
your family. It's the best way for people to get
to know about us. You can also tweet us at
Express Podcasts or write to us at podcasts at innie
Express dot com.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
CrimeLess: Hillbilly Heist
It’s 1996 in rural North Carolina, and an oddball crew makes history when they pull off America’s third largest cash heist. But it’s all downhill from there. Join host Johnny Knoxville as he unspools a wild and woolly tale about a group of regular ‘ol folks who risked it all for a chance at a better life. CrimeLess: Hillbilly Heist answers the question: what would you do with 17.3 million dollars? The answer includes diamond rings, mansions, velvet Elvis paintings, plus a run for the border, murder-for-hire-plots, and FBI busts.