Restrict, Oversee and Keep Experimenting
Zahrah Mazhar recounts the process that led to a policy on the use of AI at Dawn.com.
☰
Making AI Work for Positive Change in the Newsroom
In 1935, EB White, the author of The Elements of Style, the OG style guide for American English, wrote a short story called Irtnog in the New Yorker, critiquing the popular Reader’s Digest and other magazines that summarised long-form stories. He wrote how this led to people wanting bite-size information about everything. The result was a society without critical thinking skills. Soon, books became a string of letters. An article about how to deal with unruly children became ‘hit him’. A Hemingway novel was condensed to ‘bang!’ The news was condensed using ‘mathematical quintessence’ to a single, nonsensical word like ‘pudquitch’, which the reader could decode and know the news of the day. White predicted that readers didn’t necessarily want information and content as much as they didn’t want to feel left out.
Of course, it sounded like sci-fi in 1935, but it was an eerie premonition – think about Twitter now X. Our love for bite-sized information has not led to Irtnog but it predicted our fear of missing out, aka FOMO. And creators of information – like news media – are falling over each other to grab our attention and make dollars. But as technological innovations advance, shrink our attention spans and popular distrust of the media grows, is the future of journalism grim, especially if we are all going to be replaced by robots?Robots seem to be a popular code for AI, which was once the subject of dystopia.
However, folks may not realise the extent of AI in their lives. Whether in the form of algorithms deciding what we consume on social media, how we bank online, use navigation tools to get to our destinations and our smart devices at home, you are more reliant on AI than you realise.
Against this backdrop, no one is laughing anymore at predictions made by American computer scientist and futurist Ray Kurzweil in 2005, as many have come true or are in the early stages of development. He predicted a personal computer would fit in our pockets; he also predicted streaming devices, the cloud, pinhead cameras and even the use of drones by the military. He also foresaw what we call ChatGPT, which many of us are increasingly dependent on – even in the newsroom as a tool to assist in searches or help with writing, editing, etc. Kurzweil also predicted that technological singularity would occur by 2045. He defines it as the time when AI will surpass human intelligence; in other words, it will think better than we do, create better and even make better judgements.We are now 20 years away from the rest of Kurzweil’s predictions and while I know how AI is impacting the newsroom, I wonder what journalism will look like in 2045. My sister, for example, has not read my weekly column in Dawn since the newspaper introduced the text-to-speech tool on its site. I have listened to my pieces during my visit to her and I approve of the crisp British accent reading my work.
The Associated Press introduced automation into its workflow as early as 2016, which allowed reporters more time to add to the story or work on other stories. This is a sensible approach and frees up time for them to do the journalism, which is something no AI will be able to do because it cannot be a source or work with sources. Journalism is about trust and I don’t know how sources will open up to a chatbot, at least not at the moment. But maybe by 2045.
However, newsrooms are slowly embracing AI. In a 2024 report, the Reuters Institute for the Study of Journalism surveyed media outlets around the world and found that 16% hired a person responsible for AI-related activities in their newsrooms, while 24% said they would do so soon.
More and more newsrooms in our part of the world are using AI for jobs like automated content, analysing data (think Supreme Court judgements or government reports) and identifying trends that will help them report on issues that matter to audiences.Machine learning, (where machines learn experiences based on the data provided by humans), can be applied to many aspects of the newsroom, such as analysing large data or managing the comments section to identify problematic language.
I am especially hopeful about deep learning to fight disinformation. Deep learning, as Amazon helped me understand, “teaches computers to process data in a way that is inspired by the human brain. Deep learning models can recognise complex patterns in pictures, text, sounds and other data to produce accurate insights and predictions.” MIT designed an automated fake news detection system using deep learning models in 2019 and I hope it can be improved upon in this global era of disinformation.
Disinformation is the greatest challenge for newsrooms, and if AI can be used to help find a cure for cancer, as Kurzweil hopes, I am hopeful we will find a way to effectively fight this scourge of disinformation.
I think chatbots can also help fight disinformation and we can learn from a Brazilian fact-checking organisation’s chatbot called Fatima, launched in 2018. The chatbot helps readers verify information using social media platforms and graduated to a FatimaGPT version last year. This uses large language model technology to offer better responses to more complex issues and provide context.
But where there are threats, there are opportunities. I would love to see newsrooms in Pakistan use speech-to-text tools to translate stories from smaller provinces so they get a bigger reach in mainstream media.
Given the dismal financial state of the media, news from and about smaller provinces tends not to receive the attention it deserves. You only hear about events like natural disasters or acts of terrorism and even then, stories are reported from a faraway lens.
So many parts of Pakistan are news deserts and one way to fill this vacuum is to partner with local journalists and publish their stories in English and Urdu using automated translation tools, which will bridge the gap and help inform audiences. Work on this can happen now.
I understand fears about AI disrupting workflows and jobs in the newsroom. This is sadly inevitable when technology innovates. Some of the roles impacted included secretarial roles, data entry and bank tellers. The American service company Accenture said last year that “large language models could impact 40% of working hours across industries.”
However, it doesn’t mean humans will simply be replaced by machines, as much as they will learn how to use these new technologies to their advantage and their roles may be revised. Reporters may have more time to work on investigative stories. I do think copy editors may want to start pivoting to learning new skills to help them advance as AI technology may impact their roles. AI technology will create jobs for machine learning specialists and there will be opportunities for these roles in the newsroom too.
I see plenty of opportunities, but I still don’t know what journalism will look like. I don’t think the fundamentals of it will change – balanced reporting that incorporates diverse viewpoints, but what will change is how you receive it and in what format.
Kurzweil did get some things wrong. He thought we would all be sitting in self-driving cars by now, but that technology still faces many hurdles because AI has not mastered the skill of making judgments. I am very happy to report he was also wrong about books becoming obsolete. This is especially heartening for advocates of newspapers like myself because people want to hold books in their hands and I am pushing that example to the paper too. This is something to feel hopeful about.
Muna Khan is an instructor of journalism; X: LedeingLady
Read Comments
Related Stories
Against this backdrop, no one is laughing anymore at predictions made by American computer scientist and futurist Ray Kurzweil in 2005, as many have come true or are in the early stages of development. He predicted a personal computer would fit in our pockets; he also predicted streaming devices, the cloud, pinhead cameras and even the use of drones by the military. He also foresaw what we call ChatGPT, which many of us are increasingly dependent on – even in the newsroom as a tool to assist in searches or help with writing, editing, etc. Kurzweil also predicted that technological singularity would occur by 2045. He defines it as the time when AI will surpass human intelligence; in other words, it will think better than we do, create better and even make better judgements.We are now 20 years away from the rest of Kurzweil’s predictions and while I know how AI is impacting the newsroom, I wonder what journalism will look like in 2045. My sister, for example, has not read my weekly column in Dawn since the newspaper introduced the text-to-speech tool on its site. I have listened to my pieces during my visit to her and I approve of the crisp British accent reading my work.
The Associated Press introduced automation into its workflow as early as 2016, which allowed reporters more time to add to the story or work on other stories. This is a sensible approach and frees up time for them to do the journalism, which is something no AI will be able to do because it cannot be a source or work with sources. Journalism is about trust and I don’t know how sources will open up to a chatbot, at least not at the moment. But maybe by 2045.
However, newsrooms are slowly embracing AI. In a 2024 report, the Reuters Institute for the Study of Journalism surveyed media outlets around the world and found that 16% hired a person responsible for AI-related activities in their newsrooms, while 24% said they would do so soon.
More and more newsrooms in our part of the world are using AI for jobs like automated content, analysing data (think Supreme Court judgements or government reports) and identifying trends that will help them report on issues that matter to audiences.Machine learning, (where machines learn experiences based on the data provided by humans), can be applied to many aspects of the newsroom, such as analysing large data or managing the comments section to identify problematic language.
I am especially hopeful about deep learning to fight disinformation. Deep learning, as Amazon helped me understand, “teaches computers to process data in a way that is inspired by the human brain. Deep learning models can recognise complex patterns in pictures, text, sounds and other data to produce accurate insights and predictions.” MIT designed an automated fake news detection system using deep learning models in 2019 and I hope it can be improved upon in this global era of disinformation.
Disinformation is the greatest challenge for newsrooms, and if AI can be used to help find a cure for cancer, as Kurzweil hopes, I am hopeful we will find a way to effectively fight this scourge of disinformation.
I think chatbots can also help fight disinformation and we can learn from a Brazilian fact-checking organisation’s chatbot called Fatima, launched in 2018. The chatbot helps readers verify information using social media platforms and graduated to a FatimaGPT version last year. This uses large language model technology to offer better responses to more complex issues and provide context.
But where there are threats, there are opportunities. I would love to see newsrooms in Pakistan use speech-to-text tools to translate stories from smaller provinces so they get a bigger reach in mainstream media.
Given the dismal financial state of the media, news from and about smaller provinces tends not to receive the attention it deserves. You only hear about events like natural disasters or acts of terrorism and even then, stories are reported from a faraway lens.
So many parts of Pakistan are news deserts and one way to fill this vacuum is to partner with local journalists and publish their stories in English and Urdu using automated translation tools, which will bridge the gap and help inform audiences. Work on this can happen now.
I understand fears about AI disrupting workflows and jobs in the newsroom. This is sadly inevitable when technology innovates. Some of the roles impacted included secretarial roles, data entry and bank tellers. The American service company Accenture said last year that “large language models could impact 40% of working hours across industries.”
However, it doesn’t mean humans will simply be replaced by machines, as much as they will learn how to use these new technologies to their advantage and their roles may be revised. Reporters may have more time to work on investigative stories. I do think copy editors may want to start pivoting to learning new skills to help them advance as AI technology may impact their roles. AI technology will create jobs for machine learning specialists and there will be opportunities for these roles in the newsroom too.
I see plenty of opportunities, but I still don’t know what journalism will look like. I don’t think the fundamentals of it will change – balanced reporting that incorporates diverse viewpoints, but what will change is how you receive it and in what format.
Kurzweil did get some things wrong. He thought we would all be sitting in self-driving cars by now, but that technology still faces many hurdles because AI has not mastered the skill of making judgments. I am very happy to report he was also wrong about books becoming obsolete. This is especially heartening for advocates of newspapers like myself because people want to hold books in their hands and I am pushing that example to the paper too. This is something to feel hopeful about.
Muna Khan is an instructor of journalism; X: LedeingLady
Read Comments
Related Stories
I think chatbots can also help fight disinformation and we can learn from a Brazilian fact-checking organisation’s chatbot called Fatima, launched in 2018. The chatbot helps readers verify information using social media platforms and graduated to a FatimaGPT version last year. This uses large language model technology to offer better responses to more complex issues and provide context.
But where there are threats, there are opportunities. I would love to see newsrooms in Pakistan use speech-to-text tools to translate stories from smaller provinces so they get a bigger reach in mainstream media.
Given the dismal financial state of the media, news from and about smaller provinces tends not to receive the attention it deserves. You only hear about events like natural disasters or acts of terrorism and even then, stories are reported from a faraway lens.
So many parts of Pakistan are news deserts and one way to fill this vacuum is to partner with local journalists and publish their stories in English and Urdu using automated translation tools, which will bridge the gap and help inform audiences. Work on this can happen now.
I understand fears about AI disrupting workflows and jobs in the newsroom. This is sadly inevitable when technology innovates. Some of the roles impacted included secretarial roles, data entry and bank tellers. The American service company Accenture said last year that “large language models could impact 40% of working hours across industries.”
However, it doesn’t mean humans will simply be replaced by machines, as much as they will learn how to use these new technologies to their advantage and their roles may be revised. Reporters may have more time to work on investigative stories. I do think copy editors may want to start pivoting to learning new skills to help them advance as AI technology may impact their roles. AI technology will create jobs for machine learning specialists and there will be opportunities for these roles in the newsroom too.
I see plenty of opportunities, but I still don’t know what journalism will look like. I don’t think the fundamentals of it will change – balanced reporting that incorporates diverse viewpoints, but what will change is how you receive it and in what format.
Kurzweil did get some things wrong. He thought we would all be sitting in self-driving cars by now, but that technology still faces many hurdles because AI has not mastered the skill of making judgments. I am very happy to report he was also wrong about books becoming obsolete. This is especially heartening for advocates of newspapers like myself because people want to hold books in their hands and I am pushing that example to the paper too. This is something to feel hopeful about.
Muna Khan is an instructor of journalism; X: LedeingLady
Read Comments
Related Stories
So many parts of Pakistan are news deserts and one way to fill this vacuum is to partner with local journalists and publish their stories in English and Urdu using automated translation tools, which will bridge the gap and help inform audiences. Work on this can happen now.
I understand fears about AI disrupting workflows and jobs in the newsroom. This is sadly inevitable when technology innovates. Some of the roles impacted included secretarial roles, data entry and bank tellers. The American service company Accenture said last year that “large language models could impact 40% of working hours across industries.”
However, it doesn’t mean humans will simply be replaced by machines, as much as they will learn how to use these new technologies to their advantage and their roles may be revised. Reporters may have more time to work on investigative stories. I do think copy editors may want to start pivoting to learning new skills to help them advance as AI technology may impact their roles. AI technology will create jobs for machine learning specialists and there will be opportunities for these roles in the newsroom too.
I see plenty of opportunities, but I still don’t know what journalism will look like. I don’t think the fundamentals of it will change – balanced reporting that incorporates diverse viewpoints, but what will change is how you receive it and in what format.
Kurzweil did get some things wrong. He thought we would all be sitting in self-driving cars by now, but that technology still faces many hurdles because AI has not mastered the skill of making judgments. I am very happy to report he was also wrong about books becoming obsolete. This is especially heartening for advocates of newspapers like myself because people want to hold books in their hands and I am pushing that example to the paper too. This is something to feel hopeful about.
Muna Khan is an instructor of journalism; X: LedeingLady