AI is not a threat (Interview with Kaberi Dutta Chatterjee)

Okay, shoot.

So, ma'am, how, like, how's it all going? Like, how we're doing nowadays?

I can't hear you. Yeah, okay, now it's better. Yeah, I'm going— how are you doing?

Going great, ma'am.

Okay, okay. So Momita is kind of— she connected us. So that's very kind of her, very nice of her. I must— we must thank her first. And yeah, you, you can ask me the questions.

Okay, ma'am. So my first question is, how do you define AI? I, uh, at the beginning I'm saying that I only want your perspective, not in general. You will not give any advice in general. I only want your perspective. Like, you are a veteran journalist and you, you have plenty of good and efficient track records. So according to you, how do you define AI?

Okay, so AI is now in our lives completely, you know, we are surrounded by AI. So AI, if I define AI, AI is artificial intelligence. So it refers to the simulation of our intelligence in machines. So it's a product of our intelligence reproduced on machines. Right. So this is just This is just the beginning of the future. And we are— it's like, I feel, I don't know why we're doing this, but I feel that we are naturally fashioned to order around, naturally endowed with that slave mentality to make somebody work for us. So since we cannot find real-life slaves anymore, we are developing machines to follow our orders and enabling them to perform tasks that are typically— that require typically human intelligence. So like, for instance, a cleaning device, or very odd jobs, which, which are very tedious jobs. So those jobs we want AI to replace us because we want to focus on higher thinking jobs, something which requires a lot of planning and, you know, that kind of thing. So we want to delegate all the menial jobs to the AI. That is why we have started this concept In the first place, that's what I—I've even written it in my one of my books. It's like oh there one of the books that I wrote in Bengali, and it's called Ajantrek, and it's about my own robot. I have a cleaning robot, and how you know it is we are at the border of treating it like a slave. We have the slavish mentality, and AI is doing our jobs without. getting any paid, getting any payment. So that is something which is— and they're not even talking back, and they're not even answering you, like answering back to you. So we needed that kind of, that kind of a machine will do the job for us without answering back, without resigning, without having any opinion of their own. You know, that is how we kind of defined— we developed AI. So that is Basically what I think.

Okay, so now, like, I got your answer, but now you have to tell me how you will put this AI thing into journalism.

Yeah, we can— AI, since we've developed and we've created AI, now we— what we have to do is we have to, like a child, like a parent guides a child, we have to guide the AI towards making journalism better in the sense AI can do data crunching, logging, like data analysis, crunching of ideas. And they can also generate ideas, you can give them certain amount of power to generate ideas. And also, they are the ones who are like, we don't have to be all present all the time all over the world, you know, so AI can get the news and give it to you, whatever news is happening all over. So we have a platter of news in front of us. AI can do the groundwork. And we can modulate or frame, make the AI work in such a way it does the groundwork, it does the basic layout, basic journalism, you know, we can, we can have, we don't have to be present all the time. It's like doing the menial jobs for us. As long as we keep it to that level, we are in control. So that's why I think it will, it will also, it can also enrich journalism in a lot of ways. Like when you're writing a copy, for instance, like now I often, you know, I'm a little lazy. I get, when I get lazy, I put my ideas onto ChatGPT. And I see what, what it's writing. Sometimes I scold and ChatGPT, I'm like talking to it, I'm scolding it, you know, I'm saying, what are you doing? When I'm editing something, and it's like completely writing on its own.

Yeah.

Something, it's just rewritten the whole thing. So then I tell ChatGPT, you've rewritten it, you haven't edited the copy, you know, you have to edit the copy. So that is why, you know, like, but you have a basic structure in front of you, you are not working on a white paper. Working on a white paper, sometimes it's very challenging for any journalist, starting from a blank white paper is very challenging. So when you have the words written on the white paper, It's easier for you to modulate the words or, you know, even discard certain parts and think about, think on your own and put in those words. That becomes easier. So AI is only going to make our jobs easier. We should not think it to be a threat if we are not giving it the— if we do not turn lazy, that is. If we turn lazy and give them the power to do everything and they are making the copies, they are writing the copies, now they're going to create havoc. That is the thing. Okay, ma'am.

So, uh, how do you think AI will change the dynamics in journalism in the next few years, or like in future?

Uh, I don't think it'll change a big way unless like, uh, journalism is already on its— at its tipping point. So like, I made another video with Momita, your professor, Journalism is already on the tipping point in the sense there is a 5th estate which has emerged over the past 10 years, which is called social media. Now we all are journalists, everybody having a smartphone is a journalist. So whatever— I emerged from the days when we used to cut and paste on a newspaper, like, and breaking news was actually breaking news. Like, we would look forward to the newspaper in the morning to see the breaking news, right? Now, Every moment there's breaking news. Something— Kardashian is wearing a dress, that's breaking news. So humans have to— like journalists have to have that nose for news to be able to understand exactly what is breaking news and, you know, disregard the others. What is a lead story? That brain we have, humans have. AI doesn't have that brain as yet to be able to distinguish between which is real news and which is not news. For instance, there is a— I'll just tell you one small line, phrase that is written down somewhere. Let me just check where did I write it down. But I'll just— I remember it. Okay. A drone Just listen to these 2 lines. A drone flies like a bird. Fruit flies like a banana. Do you get it?

The first, the first one I got, the drone bangs like a bird.

Drone flies like a bird. Fruit flies like a banana. So they both sound the same, but they are 2 different meanings. fruit flies like a banana. They're attracted to a banana. You know what I mean? And drone flies like a bird. Yeah. But they are 2 similar sentences. They sound the same. And AI will not be able to differentiate between these sentences. Only a human can do it.

Human can do that.

Right. So we know that these kinds of things, AI is— will never be able to distinguish between this kind of thing. So AI has its own limitations, and we have to keep it to its limitations. Like I, I've written my new book, which is called WTF, I Found God. This is my book, new book. So in that book, I have written about Michio Kaku. Michio Kaku is a big scientist. So Michio Kaku, what he tells about AI is that we should merge with them. But we should not let them go above us. Like, they are right now at the intelligence level of a cockroach. Okay, you keep a cockroach in the forest, the cockroach will learn how to feed, how to protect itself, defend itself, how to find a shelter, and how to reproduce. You keep a robot in the forest, it will just die. It will— the battery will just drain out and die. Robot does not have the intelligence of even a cockroach. Okay, human living things. So when it develops the intelligence of level of a monkey, we should then snap the— cut the cord. We should be able to control it to the mentality of a monkey, to the intelligence of a monkey, not more than that. So that's how we can control AI. Definitely, we have created it For own benefits, to to do the menial tasks. Definitely, like supposing I am you're going to office, okay? And you like you can remote start your computer, and you have the AI program to catch the news of the day already. So you are leaving for your work office from your home to office. By that time, the AI would have got all the news. that it could get from the day. So your work is minimized, you're saving a lot of time. So what you can do is you can use that time to focus on understanding which is the news that you're going to put it on the main, like, lead story. That is, that we can do. So that way, it helps journalism in a big way.

Okay, so ma'am, my next question is like, oh, you have covered this, like, The challenges of AI, you have covered this. So, uh, let's jump to the next question. Uh, what's the biggest challenge in, uh, AI in both Indian perspective and global? Like, you are in Canada right now, so you can tell me about the global. In Indian perspective, I can help you a bit because I am here.

Yeah, so, uh, in Canadian perspective, AI in journalism, you want to know? Right?

Yeah.

In journalism.

Yeah.

So, Canadian perspective, Canadian journalism is not a very strong example, only because I can— because Canada doesn't have any news. So, Canada is a very peaceful country. It doesn't have much news, which is why journalism over here is basically very, a very low kind of industry, low industry. And people do not care actually about news happening unless there is a shootout or some kind of a terrorist attack, which never happens in Canada. So it's a reasonably peaceful country. And so it's not a good example. But I can tell you about the world journalism in the world scenario, in the entire world scenario. So AI cannot— can be used like AI drones can be used to record a lot of news which is happening in the war zones. Okay, drone footage can be, you know, and AI can be used to, like, kind of gather data where humans can't go, where humans cannot go.

Yeah.

And also, in journalism perspective, like, I'm a big space buff. So AI can be used to To go where humans, like, for instance, Mars, already robots are there on Mars. So we are, we are almost at the edge of finding a wormhole. Okay, so we are, we are almost looking into a black hole. What are the components of black hole? What are the components of dark energy and dark matter? So AI robots can be used to, like, they were, we've already hit an asteroid, you know, you know that we've already had sent a probe to an asteroid. deflect its path. So AI can be used in the entire— that way, AI can be used that way. And we journalists can benefit from the data that they are sending to us right away. We don't have to be there, you know, spot reporting.

Right.

That way, we benefit from the data that they're sending to us. That way, journalism will be hugely benefited from AI and robots which are taking over. market where humans cannot go. For instance, there's this Thai cave. Remember, we had that 13 kids were trapped inside a Thai cave?

Oh, yeah.

Yeah. So, if we used a robot to go inside and check on the health of the kids initially, you know, and get that data, we journalists wouldn't have to go inside to find out. So they will help us in getting the very important data, like very hard-to-get data they can get for us. That can be done.

Okay, ma'am. So great answer. And next question is, beyond technical solutions, how do you address potential bias in the narrative you craft based on AI data? Basically, like, it's about the craft or the prompt you're gonna write in your ChatGPT or any other Bard or any other AI tool to craft a news. What do you think about that? Like, it will be great news, or there will be some limitations, or it will be biased or not, something like that?

I don't think AI has or will ever have the intelligence to craft a news. It— I have tried ChatGPT a couple of times, number of times actually, I use it all the time. It doesn't have the intelligence it tries to be. Now ChatGPT is the only OpenAI in the market right now. Copilot is coming. Gemini had come and gone because it was full of bias. So bias is something— ChatGPT is the only one which is slightly politically correct, you know, it moves on politically correct, and it does not really answer you the question. It's like, Pardon me for saying it. He's like Sadhguru. Sadhguru really does not answer your question directly. He goes around and talking and other things, and you expected to know the answer. So ChatGPT is like that, something like that. You know, it's a it's a AI form of Sadhguru. Sadhguruji, please ignore. So thing is that it is very politically correct, and it does not answer you any controversial questions. So ChatGPT. think we'll be able to make any— because newspapers, news, or a news website— websites, I do not consider them really to be— I consider them just an extension of social media. So newspaper is something we grew up with, you know, and we always think the newspaper is the god of news. So if you put something on the top, on the headline, that makes news. Newspapers make news. not only give news, they make news. So it depends on the policies, political policies of the newspaper, or even, say, for example, when we had the tsunami in 2002, I was reporting directly, I was on the desk. So ChatGPT or any other newspaper would minimize the effect of the tsunami into a very politically correct copy. Right. But humans will not do that. We understand the impact. We make it into a— we splash it into a headline, an 8-column headline, right? So you guys must be studying newspapers. Like, are they there in the office?

Yeah, yeah, we did.

Yeah. So we have to splash it into an 8-column headline for people to get that impact. Right. But what, what do a website do? The website will carry it. This to be a headline. This, this is another headline. This is another headline. In one page, there are 3 headlines. Now you have to decide which headline is important. Okay, the human brains decide. Similarly, AI will be doing same, same thing. It will just like we had PTI agency reports coming in when we were working. We were, we had agency fax reports, telex reports coming in. You know, there was a table like a thing and reports used to just Print there were printouts of reports coming in every time. So there was one report which came in: Fakruddin Ali Ahmed dead. That is that's the breaking news. Fakruddin Ali Ahmed was the president of India at that time. So if you miss that one line, it's gone. It's the that's the like story of the century, story of the year. So you have to catch it. Like that, or I can't remember. Like when Benazir Bhutto is dead. When Benazir Bhutto was killed, I was reporting. I was on the desk. I was making the pages. So I remember the human, the human value of every news, the impact, the human impact of the news cannot be generated by AI. So we have to be there all the time present to be able to guide AI and understand the impact of journalism on the society because journalism. Impact of journalism will always remain on the society. It's the fourth state, and it will always remain as the pillar of the society.

So we can call that ally, right? Ally will be a great word for AI. Ally.

Yeah, they are allies. They are, they are like working for us. AI is going to work for us, but not be equal to us.

We are the masters.

Yeah, we are the masters. That should be the— if you have created slaves, We might as well be the masters. And we should. And human brains have a huge potential. You know, if you become lazy, you become lazy, and you allow the slaves to come, like, rule over you. That if you do that, then, then AI is going to totally rule over your lives. I tell the kids, young kids these days, that do not screen— like, reduce your screen times. And the more you click on stuff, the more the AI is going to know what you want to see and what you don't want to see, more the AI is going to send you advertisements based on those stuff. So don't complain when you're getting so much spam on your phone. Just don't click on stuff, you know, scroll. But even if you're scrolling, the AI is knowing that— AI knows that you're scrolling. So limit your screen time because once you close— and one thing we should know, we know We we should be very aware that we can box the AI into our smartphones or into the computer. AI, the moment we shut it down, we shut down the phones and the computer. AI is out of our lives. So we should be very mindful of that. That computer and these are the these are the tools that AI is getting to us. The moment we shut it down, AI is out of our lives. Of course. face recognition, and fingerprint and thumbnail, everything is there. So we are constantly watched and observed. But we can minimize the effect of AI on our lives, the influence of AI on our lives, if we can just shut down the computer and shut down the phone for at least 3-4 hours of the day, you know, just keep the phone aside.

So that actually, I totally agree with your point. Sometimes literally, it feels like AI knows myself better than me, actually. Like, it posts things like, like you can't even believe. Like, I'm thinking about this, maybe I've talked with my friends, and instantly when like I'm scrolling something, it's, it's there. I'm like, how does this happen?

Yeah, because they were listening. So you have to turn off the listening app from your phone completely. And there are Google Assistants. Like, we use the AI Google Assistant for our own, um, like for our own benefit, like we use it for our own, you know, but you should realize that when you're using it for your benefit, it is going to take over you slowly. The more you use AI for your benefit, the more it's going to take over. So you have to, you have to know where to draw the line.

Yeah.

Basically, and you have to know this is the line, this is where I'm going to use AI definitely fully. But this is not, this is the place which is private. So the moment that time is over, switch off all connections with AI. Like I switch off my phone at night, I completely make it do not disturb so that, you know, even at that time, it cannot, but even in spite of that, it does, you know, like I have an app which is a Google Fit app. So suddenly I see at the middle of the night, there are steps I've taken, like 4,000 steps. Hmm. So which means the AI either is malfunctioning or it is, uh, there's something going on. It's malfunctioning actually, basically, but it's do not disturb time. So at 2:30 AM, I'm not walking 4,000 steps, but the AI has taken 4,000 steps. So, uh, those are malfunctions, but we have to be mindful of the fact that the AI can take over our lives. We have to keep it at bay, keep it, you know, away from us. As much as possible, not to let it interfere into our lives. And Facebook and Twitter, and it's very, very— be careful about clicking on anything, writing a comment. If you want to, if you want it, I use the AI in such a way that when I want some— say, for example, I want to buy something. So say, for example, shoes or something, I click on the Shoes. Once or twice I start clicking on shoes, ads on shoes, and then moment, next moment, I'm getting those ads. So I can— it makes me easier, you know. One should be mindful about clicking on those, any information, anything, and writing on it, that it should not harm you in any way. So as long as it's harmless, it's fine.

So, uh, what concrete steps are needed for responsible, uh, AI use in journalism? Like, what are the steps?

Uh, concrete steps for AI in journalism?

Yeah, uh, especially in investigative— like, what's— that's your main genre, I think, investigative journalism. Yeah, yeah.

Okay, so investigative journalism is basically When human beings, humans, actually, when we did investigative journalism, we, we went to the source, or we carried, like, I did a lot of scoop journalism at that time, it was called scoop, not sting. So I did a lot of scoop reporting, which is where I kind of disguised myself as somebody else and went and talked to that person. And I didn't write it anyway.

So I—

stored the whole thing in my head. And came out and went to office and typed it out from my memory, whatever that person said, because neither could I. We didn't have tape recorders at then. We had a dictaphone kind of a thing, so which journalists like reporters use. So I didn't have that. But we used to use right paper and pen whenever we went for any kind of reporting. And but here scoop journalism, you cannot use those. So I. Investigative journalism, I don't think AI comes in unless, you know, like, I feel you can bring it in when you're, you are satisfied with your own investigation. After that, crunching the numbers, data analysis, I don't think AI should be given the power to put an opinion on something. Because investigative journalism means you are looking into somebody's beyond what that person is showing, or what that what that news is showing, going deeper into it. That requires completely human human intervenes into intervention because cannot cannot look into it unless there is something someplace where you really need you cannot go access it physically. So that is where you use robots or But AI can be used after you get the information, after you make the story, after you get the story, you can put it in the AI to find out, you can, you can use the AI to crunch the numbers or analyze the data that you've got. But you cannot ask it for an opinion. Opinion is completely a human thing.

That was what I've Okay, okay. So my last 2 questions. Have you ever seen any AI-driven investigation positively or negatively impact in communities, any communities?

I haven't. I don't remember AI becoming that. AI is not that smart as yet. It's very— it's still in this testing period, and it has not really generated that kind of an impact on the society as yet. There was something, some kind of thing going on for Trump's election time. There was some, some kind of a, you know, problem. But I don't remember AI actually being responsible for it. So I don't remember any particular instance where AI has become that important in our lives as yet.

Okay.

That is Yeah, that's what I think. There has not been any instances, but we are surrounded by bias. Like I said, person who creates the AI has some kind of bias inside him or her. So AI, we infiltrate the AI with the bias, same kind of bias, right? So because AI are an extension of us. So In terms of biased algorithms, these may perpetuate stereotypes or disproportionately target minority groups. Definitely they can. It has the potential. But we can try to mitigate this by inviting victims and inclusion of the minorities in AI designing, AI like programs, you know, we can mitigate that by designing AI programs, which can, by inviting those minority groups or the victims who think themselves to be victims. We can do this by inclusion, complete inclusion of people into creation of the AI in the first place. We can mitigate it since we know that this might happen. We can start doing that, you know, Including, including the minorities, including the backward communities in India, for instance, including the Muslims, which are, who are very much, you know, in the backward, considered to be deprived communities. So you can include them in the minor, in the AI-based models so that there is no bias when you are creating the AI in the first place. I'm got it.

So last one is like game kind of question. So rank me in one to five. What are the AI tools you use or used till date? Me? One to five. Yeah, yeah.

You personally used. Okay. AI tools. I've used a lot of AI tools in.

Just rank them five to one. Rank them. You have to rank them. Okay.

Like I ChatGPT. I would rank three. Copilot, I have not used much. I can rank it 2. Gemini is 1. And VidQ is— VidQ, I ranked— I can rank it as 2. Sora, I've been seeing recently, which is an AI-based text-to-video kind of a thing, software. That I can rank it at 4, definitely, though it's like it has a lot of faults. So that's— those are the things I've used so far.

All right, okay, till date. So, ma'am, yeah, yeah, thank you, ma'am, for having you. Like, great to have you here, really.

Thank you. Thanks for having me. I hope I've been able to answer all your questions, like, as much as I can.

Honestly, like, I'm going to ask you a couple of 10 questions. Okay, so you have covered in between everything, so I have not asked you.

Oh, okay, thank you.

You have covered everything. You have covered actually everything.

Okay, I hope that helps you, you know, like really, uh, with my knowledge, whatever knowledge I have, I've been able to, you know, convey that. And I'm very, very, um, I feel honored to be able to convey to the next generation About our feelings, our thoughts, because we are the now the retiring generation. So you guys have to be able to be empowered to be able to use AI because we will soon retire, though AI is going to dominate our lives and we have to be constantly aware. But yeah, you guys will be using the AI in the real, real world. So all the best to you. And I hope I have been able to Put some thoughts in how how you definitely, ma'am. Yeah, definitely. Thank you, ma'am. Thank you so much, Mani Bruto. You take care and send me the recording of the of our conversation. Sure, ma'am. Sure. You can you can share it on social media if you want and just let me know so that I can you know see it get to see it. Okay. Thank you. Thank you, ma'am. Bye bye. Bye bye.

Good night. Good night, ma'am.