https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2025 https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2025-06/Digital_News-Report_2025.pdf A presentation by our Director of Research, Richard Fletcher And Asia launch discussion https://youtu.be/0_BO65KWQPM
AI, influencers, and a public that’s losing interest: The big challenges for Indian media https://www.youtube.com/watch?v=0jo8fExS6VQ newslaundry
These are some of the biggest trends – and challenges – facing the news ecosystem today, impacting India in new and unique ways. And this was the focus of a discussion organised around the launch of the Reuters Institute Digital News Report for 2025, a comprehensive study of news consumption worldwide.
Moderated by the Reuters Institute for the Study of Journalism’s director Mitali Mukherjee, the panel comprised Nic Newman, senior research associate at the institute; Sannuta Raghu, journalism fellow at the institute and head of the AI Lab at Scroll; Ritvvij Parrikh, head of AI product at Times Internet; and Akash Banerjee, founder and host of The Deshbhakt.
According to the report, 55 percent of Indian users now consume news on YouTube, and 44 percent share news via social, messaging or email often via WhatsApp. Akash says there’s a clear indication towards television “taking a bit of a beating” while social media sees a “steady uptick”.
“You’ve got to be real and raw,” he says, and explains how consumption patterns are moving towards “snackable, biteable content”, where a shorter format is “much, much more prevalent than the longer format”.
Ritvvij talks about the challenges in using large language models that can sometimes “miss the news peg” of a story and therefore the need to deal with “detecting hallucinations”. He adds, “Trust in news is eroding and not just because of bad actors, but also because of opaque algorithms.”
Sannuta breaks down how AI can help news organisations “stay relevant” in some ways, converting text articles into videos that can be shared on Instagram. “The idea was…there are over 600 million smartphone users in India with access to cheap internet,” she says. She also explains the logistics of dealing with producing news in different languages in a country as diverse as India.
one of the differences we see in India is that traditional sources of news particularly print um but also
television remain uh really important to a lot of people especially for older groups and actually some of the print
preferences are some of the highest in our survey still so obviously that will change um but for some of the the
youngest groups we see a lot of the same trends so this preference for digital access points uh growing especially for
social and uh video networks and podcasts um uh and you know social media as a
whole is is less important but of course social media is used by all of those other people as well or most of those
other people as well as a secondary source um this chart um looks more at volume of
social media and you can see how extensive uh different the use of
different social and video networks is in India overall across across the whole sample so over half of our survey
samples say they use YouTube for example for news for news specifically and obviously more for other purposes around
four in 10 use WhatsApp weekly again which one of the highest uh figures that
we see in the world is a particular characteristic of the Indian market sharing of stories whether that's by
social media or messaging apps also you know incredibly high around 44% of our weekly sample
and just to put that in perspective let's look at this across a whole range of different countries YouTube in India
for news uh as high as anywhere in the world so uh only Thailand has similar levels over 55%
um uh and that's you know you compare that with with the United States where we've already seen the influence uh 30%
so proportionately higher high proportion of people using YouTube for news in our survey we've also been able
to identify the names of some of these influencers and creators in in all of the countries in our survey including
India uh and we asked people to name uh those individuals that they followed in
networks like YouTube for example and these are some of the most widely mentioned including I'm pleased to say
one of our um panelists today Akash um the figures by the way are the number of
followers um for each of these in in YouTube uh so what's striking about this is that many of them are men of course
some of them are former journalists like Ravishkumar uh others are younger creators um uh Drew Varafi for example
uh operating out of Germany uh it's also important to note that in countries like
India many of these creators are off able to offer different perspectives perspectives that are critical of the
government so this is maybe a different dynamic to what we see in the US where many of those creators are more
supportive of Donald Trump for example as well as political commentary we have
news adjacent figures so uh sort of comedians who also sometimes talk about politics we have satire we have younger
YouTubers uh creating content this husband wife team for example creating content uh for Gen Z uh and we also have
politicians building direct connections you know with huge numbers of followers uh Narendra Modi but also businessmen
like um like like Elon Musk as well um big followings uh directly and
indirectly and part of this is happening of course because of the growth of online video
and part of that is driven by just less friction uh lower bandwidth costs etc
one question we ask uh fairly regularly is is whether people prefer in different
countries to read the news to watch the news or to listen to the news and it's
really interesting to see very significant country differences so firstly
um when you ask across the whole sample we see that most people still prefer to read so that's the orange 55%
uh in some countries like Norway and the UK that those numbers are even higher so around 70% so they prefer to read so
this is about the convenience when you're active looking for content looking for content then reading is
really really convenient but in some other countries like India uh the Philippines and Mexico we actually find
that more people say they prefer to watch than prefer to read so again I think this really helps to explain why
YouTube and online influencers seem to be playing a bigger role in India than they do in many other countries moving
on to another sort of huge topic uh that of misinformation and disinformation so
overall uh around half of our sample in India worry that what that they can't
tell the difference between what's true and false on the internet when it comes to online news but this year we wanted
to get a sense of how people check if they suspect something really important
might be false you know where do they go do they go to YouTubers do they go to chat bots do they go to traditional
media so first thing to say is what people say in surveys may not be what they do in practice but what we find
overall is is a relatively high and clear hierarchy of intent so uh at the
top there new sources I trust this is across all countries uh along with
official sources so there's something here about reputation and heritage that still matters even if people don't go as
often if we look however at the results from India in the pink we see some
interesting differences so much greater emphasis for example of using social media as a way of checking uh or
comments from other users um or somebody I know personally that
could be a relative it could be could be a friend so I think this idea of peer-to-peer trust rather than
institutional trust seems to be relatively more important in India another interesting difference perhaps
to pick up in the panel we asked people to name the brands they would use to check so which brands did they see as
trustworthy in this context and it tends to be in different countries uh traditional brands quite often broadcast
brands with a record for uh breaking news so in India uh BBC and ETV times of
India as well with our more educated sample remember the more educated sample um brand recognition for factecking
brands a lot lower um but it is interesting to see that as alongside sort of factchecking organizations we
also have traditional media organizations like BBC verify times of India fact check for example also
showing up there and when it comes to who the public thinks is most responsible for spreading misinformation
we find at the top uh politicians uh and online influencers as well which is
interesting But it's also striking that almost four in 10 think that the news media and
journalists are also responsible for spreading misinformation or false narratives which by the way is a little
higher than uh many other countries in our survey and so clearly there's a problem of trust in journalism
linked to in many cases perceptions of undue influence by powerful politicians or powerful businessmen
so given that uh and by the way India is is is is not alone in this what what can
the media do to increase trust so trust in India roughly at the moment uh 43% so
about four four in 10 say that they they trust the news so this was this was one of the key
questions we asked our respondents this year we've coded the responses into different themes so the first thing that
the public thinks uh the media could do is around showing greater impartiality
or objectivity people worry that the media is full of biases it has a lot of hidden agendas uh often shows this
through loaded language sensationalism in different ways and in surveys they
say they want the facts so they can make up their own minds secondly uh we have a cluster of comments about accuracy about
verification the importance of verification again it's interesting to see a number of publishers really sort of doubling down on a lot of these
themes i've mentioned that already greater transparency is also widely mentioned people say they want want the
media to show more of their workings they want better labeling of things like conflicts of interest i think that links
to the impartiality point at the top and then finally better journalism so more
deeply reported less clickbait less sensationalism uh holding power to account um so a
pretty consistent picture in line actually with what many journalists would want to do but of course this is
not what consumers think is currently happening or many consumers which is why we have this sort of disconnect between
uh the perception and the reality so a final big headline from the report
is the emergence of AI chat bots as a source of news and I should say this is
quite hard to ask about in surveys because sometimes consumers aren't aware that what they're seeing is necessarily
generated by AI for example uh Google AI overviews overall across markets we find
that 7% so it's not huge numbers but you know significant numbers say they've used a chat but chat bbot to ask to um
access some kind of news in the last week this figure is twice as high in India but again remember that sample is
a bit more educated and English speakaking as we discussed last year um
people in uh most countries are deeply uncomfortable with the use of AI in the
news particularly in situations where AI itself is creating the content with only minimal human checks so look at that red
bar in the United States 53% uncomfortable but it's a really interesting and different story in India
so only 20 20% uncomfortable 44% said they're comfortable with this kind of use of AI
and then in the sort of other case which is where journalists are using it in supportive roles so it might be
transcription or backend uh support in tagging these kind of things um there is
much more comfort both in the US uh and India for example
um it's really interesting that um this sort of greater comfort with the use of AI and used isn't just in India across
many countries in Asia in particular we see this we also see a lot of experimentation a lot more
experimentation actually than we see in Europe with AI news readers here's one from Thailand social media avatars
translation many other functions that are much more audience facing we see a lot more caution in in um and reticence
in parts of of of Europe so just 11% are comfortable with this use of AI in the
UK for example in terms of what lies behind these levels of discomfort we ask respondents
in the survey how they felt AI would change the news in the future and across countries people feel it will make the
news cheaper it will make it more abundant effectively it will make it more upto-date but at the same time they
think it will make the news less transparent less accurate and crucially
less trustworthy so potentially I think there's an opportunity for news organizations to emphasize these
elements in their reporting which we've already heard are really important to audiences
again it is a different story in in India so people more like to see the upside across a whole range of
attributes um including I think this one is really interesting can AI be used to
make the news easier to understand and there's a lot of work being going on which I'm sure we'll talk about in the
in the panel about trying to make news more accessible in different ways and we
actually asked people in the survey about this to tell us about these sort of different ways in which AI could be
used for transforming things and you can see on the left there in India sort of high interest in a lot of these things
uh right at the top different reading levels so maybe making news simpler or changing the words for example that's
much higher than in any other country that we uh that we surveyed um uh so
also summarization article translation India has many different languages so I think a lot of potential for using this
technology to increase accessibility and relevance and it's really interesting to see that there's a lot of interest in that so a lot else in the report I
haven't even mentioned things like use avoidance yet but I'm sure we can get on to that uh but just quickly to recap um
so there's this sort of trend not in every country but the media is losing influence as politicians audiences are
tapping into this sort of alternative ecosystem so that's really challenging in lots of different ways for media
companies how should they respond to this trend what does it mean for talent video and audio formats becoming more
important uh but text also remains important it's not really an eitheror but but how should again you know should
more more resources being shifted into uh video content for example trust is
relatively stable now um after sort of declines um but you know what can we do about
some of those factors there's clearly still a disconnect um how could that be improved and then that challenge for
challenges and opportunities for AI chatbots and really interesting to hear what the panel says about those the full
report of course is on our website uh lots more in it please do go and dig in
i will stop there and hand back to Matani nick thanks very much for that and in many ways um India is hugely
important because some of these big themes that we're seeing globally whether it is sort of the expanding of the news ecosystem or what's happening
with AI is really being reflected for India so let's open up to our panel today akash first round with you it's
you're someone who sort of transitioned from traditional journalism into being in social media networks as of the 16th
of June your subscriber count is at 5.76 million so it's been sort of ticking higher and you've chosen satire as a
medium to speak about you know issues that are not always let's say lending
themselves towards humor tell us what your experience has been in terms of shifting to some of these ecosystems why
you did it uh and what you think of when you're engaging with your audience uh so
that I I think the headline numbers very clearly indicating towards the fact that uh television is taking a bit of a
beating and and and and social media uh has been seeing that steady uptick and and this this is something that we have
experienced um the the interesting thing to add here is that 7 years ago when we
started the dish bak the whole idea was political satire the late night American television format nothing revolutionary
uh in its own idea but revolutionary because India would have hundreds of news channels and we didn't see news
satire or political satire at all and I was like why is that uh and well one comes to realize that if you start doing
Steven Colbear Trevan O in India they'll be in jail the very next day uh because that sort of political satire isn't
taken very easily uh it is seen as personal insult uh but we wanted to experiment with that on the internet
what is very interesting is over the years we have constantly been getting the feedback from people is that while
your attempts at satire is great but we are not getting the news from mainstream media please get us the information the
raw information u the non corrupted mainstream media information so we've
actually pivoted a little bit and given people the information that they want with of course a sprinkling of satire so
what your research is showing is what we are living on a daily basis where at least the urban audience is shifting
away from mainstream media however mainstream media has the resources the depth it's it's still very very big uh
for us digital folks to make a complete impact all over the nation but the trend is very clear vitali it's also akash
clearly a market where video works really well and that probably sort of bridges you know an urban rural um
access into information um walk us through what your experience has been in terms of using video but also a trend
that we're seeing worldwide now is sort of these podcasts which is a podcast but it's also recorded on video it's
consumed on a platform like YouTube have you have you for into that space as well
um so we do video podcasts uh they are longer format um I I think one issue
that I may pick here and and that's al also being pointed out by Nick is that you talking about an English audience
who has given you this feedback what we are also seeing as a real world feedback
is either there is news avoidance or there is this whole fatigue from very
long format news india is not a long format news consumer and we've seen that in print also everybody wants bite-sized
information we've had people come and tell us "Oh we used to watch you but what now we don't?" I said "Why what has happened are we suddenly so bad?" They
said "No you know 15 20 minutes is too long we are more interested in watching shorts so one thing that we are seeing
continuously is this whole evolution the whole Tik Tok uh generation you might
just say we might sound antiquated to do 10 to 15 minutes worth of content where everybody wants to consume content even
political content in one or two minutes does it drum it down yes u can we scream
and shout about it yeah maybe but that's not going to change the consumption pattern so the consumption pattern has
gone towards shorter crisper content the video podcasts that you're talking about
the the popularity of those podcasts are actually not based on the entire 1 hour video they are based on the clips that
are taken out of that 1 hour video and made into one or two minute snackable bitable content so the shorter form of
video is much much more prevalent at this point of time than longer format rit it's a great point to get you into
this conversation what are you seeing in terms of sort of more me mainstream trends uh for print for you know online
digital audiences and just on this point about video what's also clear is that most of this video is being consumed via
platforms not the website itself what's the experience been for the times group and how are you approaching reaching
more audiences so I agree with the fact that uh postgenerative AI we are able to produce
the same news but in different forms for different audiences Additionally uh we we are historically a
print first organization and it is allowing us to produce more video at a
larger volume so it's allowing allowing us to create content for both our apps
make it more userfriendly and uh accessible to people are you exploring
doing more on these platforms though Rit because you do have that unique advantage and in that India you know
stands out amongst others where media organizations like the times of India actually hold as much clout in terms of
you know reaching audiences as some of these social media platforms to are you needing to choose between one or the
other we publish on both uh we've increased our YouTube production drastically over
the last couple of years which wasn't the case I think even two or 3 years ago uh but the focus also remains on how do
we make our apps much more uh attractive to audiences so that we have much more first party engagement directly on our
app itself we didn't touch upon news avoidance in the main presentation which
but that is a trend that we've seen for two years running now extremely high numbers almost a third of you know the
audience comes back saying it's too overwhelming it it makes us feel worse about things and we tend not to engage
with it um India has been through a pretty intense uh news cycle starting
off with the elections last year and then increasingly through this year there's been a lot of sort of political
defense and other sort of news what's your experience been what are you seeing with audiences both on the print side
and your online engagement so one of our strengths is that we have reportage across the country and we have very
strong city/reional coverage however uh the traditional news homepage never
surfaced these stories and as we adopt more personalization algorithms the
consumption of stories actually spreading out from Delhi focused
mainstream content towards what people want and which is much more regional focused so uh I think people come to our
apps because we are able to showcase information which is much more divergent from what is there on the mainstream
homepage senator in many ways our sample audience sits quite close to scrolls
target audience so you know it would be higher educated very comfortable with using and consuming news online and
what's the way in bin in terms of you know broadening this out and reaching more people
yeah we did an interesting exercise you know in um in in 2019 which was so we
had just again we had gone through you know the paper two video i had started the video team at scroll uh it wasn't
sustainable uh we had to let go of a you know a big chunk of my team and uh which
is when uh I started looking at how sort of AI could could uh you know help us uh
be relevant in this space right so we did an experiment uh uh I mean which got
funded by Google and uh it was basically are we able to convert our text articles
into uh into videos that could be shared on Instagram uh the idea that at that
time was you know there were 600 million smartphone users in India with you know access to cheap internet uh and scrolls
audience was uh you know the 30 million English universe uh are we able to uh
sort of make a dent in the wider circle that is available to us uh and a lot of
the things that Nick mentioned uh you know with regard to AI we were we were very very early in those experiments but
a lot of the things that Nick mentioned hold true which is uh you know uh in in
a lot of our audiences didn't really have uh that much of a problem in fact we didn't receive a single comment uh
saying you know this is fake this is AI generated you know put this away uh and
and we saw sort of great uptake on it and also it increased our uh you know
followers engagement and all of that and we did this during the election uh in in 2024 so it was an interesting experiment
uh which which did work for us and uh yeah
you've segueed into AI so that's great for me sonu let's talk a little bit about you know some of the findings in
terms of where interest is highest because interestingly it's also the most challenging things like uh language
translation for instance in India is hugely difficult also in part because of what our spoken language is like we may
be speaking two or three or sometimes four languages in in a couple of sentences um has you know has crew been
thinking about language translation and tell us more uh yes we have been
thinking about language translation but it's a very difficult problem to solve and I've I think I've spoken about this
so much at this point uh you know the government in fact is has started this bhashani program so one would like one
can only imagine how massive a task this is uh for a country like India with 22 official languages uh you know and and
to make it accessible uh in the smallest of dialects in the smallest of accents
uh to to people across the country it's a very very massive task uh what we have
right now is uh you know organizations like Serbam for example who's who's been
uh sort of you know commissioned to build India's sovereign uh LLM and uh
those those are available for news organizations uh you know to to take and
experiment with there's of course multiple models and multiple uh data sets available on the Bhashini platform
for us to experiment with but what we're seeing is so again we we've tried this for about uh 3 years now and uh the the
you know the the way it's improved has been fantastic what it was in 2022
versus what it is now uh the improvement has been uh you know very very large uh
but are we able to use it in an accurate manner are we able to use it uh you know
with fidelity to source uh I would say not yet uh even within that language
pool I think Hindi does uh best and you know I mean this can primarily be uh
because the data sets of Hindi are much richer than say a conany uh you know or a methile but uh it's it's I think it's
going to be a long process but we are we are doing as a country I think we're
doing a good job rit you've been leading some of this you know this AI charge uh
at your group for many years now In television it looks like there are experiments such as an AI news anchor i
know they're doing that in local language in the southern cities as well what are you experimenting with at this
point at uh the times of India in terms of AI and you know engaging with audiences using that there are
initiatives across the board i don't know how to recollect all of them uh here's the key highlights
but I think one of the key things that we've realized is that LLMs are essentially summarization systems and
they often miss the news peg so even within English to English rewriting
it'll miss the peg of the story so it'll try to summarize it to whatever is the largest part of the story but what was
the original intent it might miss that additionally uh uh there's a great
amount of work that's happened in the company with respect to identifying hallucinations between using of LLMs
so uh even if we use it it does not uh miss the point or change the story uh so
while I I think the use of LMS will result in lots of applications infrastructurally I think detecting
hallucinations is one of the big things that uh we are working on just quickly
Ritu is the thrust of the focus though on sort of you know how to use AI in terms of changing practices um and ways
of working within the newsroom or are you thinking more deeply about the audience connect and as an extension of
that um any thoughts on the labeling debate no uh I think both uh operationally of
course it has helped editorially editorial teams ship out more content but also from audience facing
perspective I think we are able to now generate uh AI anchor content videos with AI anchors uh create crosswords
with the latest news in it u uh then creates snackable there's an Instagram
channel which is not officially times of India right now but uh we're testing out content which is simply visual doodles
and stories but the latest news as doodles and stories so there's a significant amount of format translation
that's happening but uh infrastructurally again finding the the the newsp adhering to it ensuring that
hallucination does not happen if we crack the infrastructural layer each team and each editorial team might want
to come up with their own style and their own way of doing things nick we were chatting over the weekend and you
mentioned how it was quite interesting finding and looking at this AI data for India specifically is it quite different
from the rest of Asia or does it feel like a trend across Asia where there's greater openness to sort of
experimenting with and engaging with AI in in the new genre absolutely i mean I
I think we've just heard you know some of these examples are are real and live and they're audience facing it's not
just about efficiency in in place like India um so and I think that comes from the sort of greater willingness to
embrace um technology and and sort of uh the the upsides if you like of of this
AI revolution but I I you know I also think there there is no sort of silver bullet this is going to take time as
Senut was saying um and uh and I think there's a lot of sort of we haven't really talked about the access model
part of this you know so I think whatever we're hearing about a lot of attempts to create more content in
different formats for example but um there's only going to be limited attention and the access points are
still going to be choke points whether that is a search engine or a chatbot or whatever it is who what is going to be
surfaced of this incredible abundance of of media that that is that is about to happen to us akash are you using it or
do you hear anecdotally of the news influencer ecosystem beginning to use AI
because it's also potentially a tool for you know new structures that are not
that uh that are not that financially well padded um so let's get a few things
completely uh out of the way first and foremost uh television versus social media we now know the trends are there
and and television is going to have a tough time um and social media is gradually going to win the race so we
know that is AI going to be used in the back end absolutely we have been using it for some time now for example
subtitling earlier we used to have a boy who used to sit has used to listen the whole Hindi episode then translate into
English now we use AI about 95% accuracy listens in Hindi gives me English
subtitles our subtitles get up within half an hour 45 minutes of an episode going so that has reduced the time taken
from hours to less than an hour as far as subtitles are concerned however I would tend to disagree with the other
members of the panel as far as uh use of Gen AI in terms of delivering news uh
and some of it is anecdotal uh one of the newspersons that you mentioned in your report did an experiment where uh
an expensive studio was hired i'm not even talking about a language LLMs um
and the voice was translated by industry movie level dubbing
and uh parallel channels were launched and they failed miserably
and I I think there is a reason why say Malali for example you start you've been in the world of business news anchoring
imagine you starting to talk about the stock market in Hindi uh in in absolutely pure uh Hindi would you get
that sort of attraction because We might use these LLM models in education
absolutely a and that's great for the for the country but I think where this issue of trust comes it's the same story
why people are not listening to television at least the educated ones are not listening to television and going to social media is because there
is a trust issue the same thing I think will happen is even if you get and right now it's a bit of a joke youtube
recently allowed for audio automatic audio dubbing therefore my language can be available to somebody who's speaking
Spanish and it will automatically do it it's a bit of a joke even English is a bit of a joke it is it is very monotone
it is a little bit of a comedy uh it will never get those inonations right so there is as Nick was saying a lot of
work that remains to be done but even if it is done ICE and and my livelihood
will depend on it because otherwise some big company will come and do a gen AI of Akash and and and give all the
information and the satire i believe that with AI still coming in you will
have that issue of trust when the user will get to know what I am looking at is
an AI model that person might have a bit of a trust issue right now I think people don't
know what AI is and what it is capable of they are listening to these AI models and bots once the awareness will
increase I think the genuine touch and contact will be missed ai will be a lot more in the background in research in
scripting in subtitling but that human touch uh I I think will continue to be
very very important as far as news is concerned you've picked the the prickly item so let's uh let's sort of address
it which is trust our cash and sort of it's two sides of a coin one is the fact that trust has been stable but it's not
high at 43% and the other that for India is a market misinformation remains a
really really big concern what we've also seen is that channels like WhatsApp are used most frequently even if they
you know the discomfort to them is pointed um what's the environment like
uh considering the sort of the news cycle India has been through there were the elections which were quite fraught
as we know we've been through conflict between India and Pakistan and all of these have seemed to be peak events in
terms of the flow of misinformation particularly on social media platforms
this is our bestselling t-shirt on the Desh Bhta it's called WhatsApp University i didn't think that where and
the bottom line is where facts are not facts uh and India of course has this particular problem as far as WhatsApp is
concerned because it seems that because it is written it doesn't matter whether it's written on digital because it is written it
might be the gospel and and it's it's entire generations i mean from youngsters to aged educated people we
all have this in our families we all know of these stories uh it is an epidemic level uh of a problem so you
have mainstream media that is compromised you have WhatsApp which people will believe in blindly uh and
that's why WhatsApp has put so many different things about not being able to forward more than five people etc etc even WhatsApp knows that this is a
problem of overbelief um I think what what is happening is organizations like
us survive with a very small team is because there is still that part of population that will come to you seek
you out want the genuine information what one can hope is as the technology
evolves remember this whole MMS era where you had these secretly filmed videos in India that used to create
scandal i think what has happened traditionally in our country is technology will get misused first then
we'll get to know how to use it properly my hope and I can't give you statistical information on this my hope is that as
technology evolves as people get used to it uh they will start to have this thing
about trusting what sources they listen to we are seeing a little bit of that in urban India but the problem is is quite
big and and and let me tell you uh influencers are as as guilty you have as many fraudulent influencers also i do
not speak on behalf of the influencers as a matter of fact this is the next level of problem where you have
financial information uh you have these uh influencers who are selling stock market tips which are absolutely
fraudulent and the SEBI is cracking down on them so you have legal misinformation
political misinformation financial misinformation and for a country like India this is huge because data is so
cheap and everybody is consuming this for those that may not be across this Sebie is uh the securities exchange
board in India and there has been sort of a blurry line between news lifestyle humor investment all of that merging
into news rit come in on this point about misinformation uh Meta's largest investment in terms of supporting
fact-checking organizations is in India times fact check you know is an organization or a vertical that focuses
on it has it changed has the dial changed or has it just always been you know quite intense
i live in a very sanitized environment within the newsroom so uh you can't say that you can't be working in India and
say that but uh the focus constantly is to ship
out find facts ship them and uh like so
actually before misinformation I think we so there no one will nothing
can compete against human touch i agree to that but the pace of change of
technology is so fast that it is inevitable that a lot of production will move towards uh genai and the
differentiation that news companies will continue to hold is news gathering and news verification you crack these two
things everything else will slowly move towards technology so I'm a extreme and
Sar has heard me say this since last I don't know two years now this is the only uh mo that we hold otherwise
Facebook can create information Google can create information news gathering and verification continues to remain the
primary uh mode that we hold uh now coming to misinformation like
once you've gathered the news structuring the entire facts in a way that it can be shipped in different forms so that it can actually be
packaged into uh you know uh tablets of fact on Instagram or on shorts or any
form but knowing which facts are important and which are not is again information processing that needs to be
um if I if I can give a technical example um if you look at Google sheets Right uh a Google sheet is essentially
data plus formatting that helps us as humans understand information there'll
be colors there'll be pivots and there'll be uh stuff like that uh but if you look at a CSV which is the cleaner
form of a Google sheet and you download it there's no human formatting in it so
separating out the the information from the style and the tone and the all of
that so that the raw information is saved in a way that LLMs can further
send it to audiences in the right way without hallucination is possibly the key to tapping into your archive in the
future sanu as a smaller news organization but an organization that has sort of focused on on both these uh
levers if you will how are you thinking this through I mean scroll in the past has done several explainers particularly
around large events and you're pivoting quite hard into bringing AI into you
know not just your newsroom process but how you're engaging with audiences is that something you're trying to merge
multiple points and I'll start by sort of addressing yours first which is so we are a small newsroom right and we've
always used technology as a force multiplier so one of the things that we're building now and sort of uh 2.2
two of the text to video that we built uh is a text to interface tool so the
idea here is that uh you know say in in 3 years time 4 years time uh 5 years
time uh users may have the option to pick the form that they want to consume
the information is and uh you know rituage was was bang on in saying that
our mo right now is essentially a verification i would I would go as even
far as to say that you know even use gathering can be done uh you know by by machines in the sense of you know look
for uh stories in large data sets and uh you know pull them in of course going
out and getting a story uh will be the the moat that we have that that we have
and of course the verification angle uh so what we're what we're saying is are
we able to play with a form uh that can be dynamic so our journalist produces uh
you know a verified reported story uh at at one end and then we are able to
provide it to a user in a form that they want now this form could be uh three bullet points they could be a podcast it
could be a calculator right so calculator is something that we're really really sort of uh you know
experimenting with i mean the budget is a big festival in India there are so many stories uh that that lend
themselves to uh be uh sort of served as a calculator as opposed to this wall of
text which is 500 words long saying the finance minister said this these are your tax labs uh there are other things
that you can do with calculators the other sort of uh format that we're trying is something called a context
slider which is uh again one of the points Nick mentioned is how do you make
this easier to understand for users right so we're trying our our sort of primary assumption is that uh every user
has a different entry point into a story why are we then uh you know giving them
this 500word uh block of text or one single video which is not linear which
needs needs sort of uh you know navigating through uh so we've developed this context slider where if you're a
beginner it gives you uh information in uh you know just the basic ones if
you're deep into it it gives you information uh at at at whatever level uh that you want to consume the
information at right uh so I think uh these experiments are uh important we've
done some very early AB testing uh you know there's there's there's good uptake in it and you know to Akash's point
about uh you know the the the the anchors in Genai I think there are multiple manifestations of of geni and
that's what we're seeing right like all of these things like the calculator the slider uh we're we are experimenting
with a bunch of other formats they are all manifestations of geni and we wouldn't be able to do it without them
how do we how do we code journalistic intuitiveness into the things that we're
doing how do we code journal istic logic into uh these products that we're building and that would then go ahead
and uh signal trust right and how do we have fidelity to source so we're we're
very deeply thinking about these things uh you know when when we're building these products how often do you tap into
user feedback Sanu and you know what's the best way for you as an organization to source that uh we look at everywhere
in the sense mostly I think what people comment on your work in a in a public
forum uh is is sort of something that we tap into a lot so another example is you
know when we did put out videos that were generated from our uh text to video tool right uh we spent a lot of time
optimizing on an avatar so our one of our earlier uh very very early policies was that we wouldn't do human sort of
photorealistic avatars uh so we wouldn't create an Akash and put him on our uh
you know on our video for example um there's only one Akash Sanu so there's
no way thank God thank God there's no way we that like Narish was very
specific about you know Narish Fernandis our editor was very specific about you know we will not touch the
photorealistic space at all uh so we then optimized uh with sort of uh
illustrated avatars and the first video that we put out people hated it right again you know they said you know what
is this we don't expect this from scroll this is you know absolute rubbish you know we we want like a whole lot of
comments which came very very quickly then we did the second experiment where
we said you know Indian in Indian Instagram is very scrappy right we we've seen the Ganji chels and the you know
multiples sort of manifestations of the scrappy Indian Instagram right so we said why don't we do this in a scrappy
way so that it may help people connect with it better again it was it was it
was a guess it was an assumption that did like a little better but people were still saying why are we seeing this from
scroll this is you know this is a little silly like you we don't expect this from you uh again the comments were like
almost the day we put out the video there was no lag really but when we did move to a clean UI when we did move to
like better looking fonts cleaner you know colors and things like that not a single comment and people were engaging
with uh you know the the the the content and what we were trying to say uh so
it's it's it's it's definitely sort of interesting to to look at it uh uh you
know as an experiment and say uh you know this this is working and this is not this is uh people are responding to
what we're saying and uh and and yeah and and I think that would be our our
way to to look at user feedback one of the ways at least one of the things we
haven't touched upon is sort of subscriptions which is in many ways the bread and butter of news businesses and
for a majority of our survey respondents nothing will you know get them to sort of pay to subscribe to news but having
said that some of the numbers in India are quite eye popping the press Gazette's June data for example points
to the top three fastest growing sites in English language and that's the Hindustan times the the India times site
and india.com are you you know confident that you continue to see good traction
in terms of audiences coming and engaging with the news because ultimately you know that's the frustration and that is the end goal
increasingly in TI and ET people are buying subscriptions the transition is happening uh
it's taken us a lot of time to identify what is the right content that they'll pay for uh so uh I think things that are
more utility focused uh drive subscriptions strongly additionally uh instead of I think as a
as a organization we don't see it only as subscriptions we see it as reader revenue so people might pay for events
people might pay for classes like courses uh they might pay for completely
different things till we can get direct revenue from audiences that falls under one bucket which is reader revenue and
different products appeal to different people um yeah Nick is that something
you saw while tracking the data across different countries is India um an anomaly or is it sort of the higher end
in terms of appetite to to consume news and to subscribe to it uh yeah I mean I
think uh subscription clearly is working for sort of really big upmarket uh news
organizations that are um talking to richer educated people i we've seen that in the US and many other countries um um
and it also is working for where you can provide value in a particular niche as well so I think both of those substacks
or or areas where you know the information is really valuable it could be technology finance etc so that's
really where we're saying subscription but that's clearly not going to work everywhere and in a lot of countries outside sort of northern Europe and and
and the US subscription is really really hard so I think as Richard says it's a lot of this is about revenue
diversification and understanding you know where you can find utility uh and and and definitely how you can bundle
things together so people are you know that's the real trend we're seeing right now so if you can bundle courses games
you know with news then people have to think three or four times about unsubscribing from that so so this is
what the New York Times has found and others so I I think that's the big trend um and there's definitely no silver
bullet again you know this is this is really really hard and it's about a combination of different things akash
final word with you and in many ways the elephant in the room which is the the
media freedom background that India is currently or Indian news organizations operate in the sort of level of business
ownership of media organizations why did you choose Akash to move from traditional news into building something
of your own and having built it do you feel stronger or do you feel more
vulnerable uh as a news journalist who sort of present as an independent entity on in on an another website or another
platform so um I I I think in in the future uh uh this this I'll just start
with this because this has been bugging me since uh Nick uh uh told us about the opinion and not facts and journalism is
it really happening and this is what the audience is really asking for um this needs to have a deeper I mean this is
also the elephant in the room um you know we we constantly at least few journalists try to uphold themselves to
the highest caliber um let's not be partial let's not take money from here i mean I can I can say in the seven years
that we have existed we've never taken one grant and one political funding or one dubious source of funding and that
that's all nice and good but does the audience honestly care does the audience really want facts really the top
channels in India are all opinion based news channels are all about 9:00 debates are all about opinion so somewhere down
the line there is a crisis in the audience and the crisis in the readership as well and we have to realize that and we have to then craft
uh whatever we are offering we are completely unashamedly we say this we exist as long as people are supporting
us are viewing us there is ad revenue uh we will stop being so uh you know so
balanced uh if if the audience doesn't care and I I say it on their face uh are
we doing the most important stories that need to be told no we are doing stories that need to be told but people also
want to watch so I am not going to do climate change episodes because they will tank miserably because nobody cares
about environment change in India as much if you look at the larger audience so one has to be a little bit more
realistic having said that that there is a little bit of disappointment as far as the audience is concerned in this
country um the liberating force that a small media channel can give you a
social media presence can give you the recognition that it can give you the freedom that it can give you is nothing
compared to 10 years of television uh you've been in the world of television and I can just tell you uh the
difference the stark difference of 10 years of doing television and then doing one or two years I'm not even talking about seven years one or two years of
doing digital people will come up to you they will know you by name they know exactly what you're doing so the kind of
hold that digital creates the kind of hold that independent content creators are making is far deeper than any news
channel mainstream channel guy could ever hope for the trust that they have is far deeper than any mainstream guy
could hope for now the question is will they be able to survive there is financial there is legal and now there
are procedural issues of content strikes coming your way the challenges are humongous but the freedom that this has
given and that's why you're seeing so many of former news people coming onto YouTube coming onto social media platforms in India is because it is such
a liberating force fair point Akash and um actually not just in the digital news report when but in series of our
research when we talk about the concept of transparency audiences often point to the why rather than the how they so much
to lift the hood of journalism and understand the process of a story they want to understand the motive behind it
so maybe sort of food for thought for news organizations and can I say I'm very glad I'm not on television anymore
i'm like a dinosaur who is happy to have been in the golden age of TV when it was but thank you all very much for joining
today it's been a fascinating conversation nick Newman lead author of the digital news report um as he said