Artificial Intelligence is the buzzword these days and news media is no exception. So obviously, AI in news media was at the center of discussions at the digital INMA South Asia Media Summit held on 10 and 11 August 2023.
In an interesting presentation on ‘How AI will transform the production, consumption and monetization of news’ Dr Shailesh Kumar, chief data scientist at the Centre of Excellence in Al/ML, Reliance Jio, India, quoted Alphabet and Google CEO Sundar Pichai – “AI is probably the most important thing humanity has ever worked on. I think of it as something more profound than electricity or fire.”
Kumar started by saying that the world is at a very interesting junction where two powerful forces – AI and the media – are coming together. He shared examples of how AI is already starting to transform the sector – obviously with its pros and cons – and a few workflows.
The magic of tech
The convergence of technologies such as the internet, connectivity, devices, cloud, internet of things, and machine learning is creating magic – without which we wouldn’t have cabs, food or news apps, or even ChatGPT. “It is also time to think about how the evolution of technology is affecting the evolution of businesses,” Kumar said, recalling how Pichai’s statement didn’t make sense to a lot of people when he said it but we are, nevertheless, having to grapple with this new reality today.
In the last 20 years, AI has affected only a few use cases such as building a search engine, technology such as ChatGPT or Netflix or Amazon recommendations, or in select areas such as printing technologies. But apparently, there has been an explosion in the use of AI in many other social care sectors such as agriculture, education, and healthcare.
Artificial intelligence, however, comes with great power and as such requires great responsibility too. “The question is not if AI will take over but if it can be used responsibly or not,” he said.
Coming to news media, Kumar outlined the evolution of news over thousands of years – from oral storytelling within communities to manuscripts and the mass mechanical reproduction of text, images, sound, and video using tools and media such as the printing press, newspapers, radio, TV, digital, and social media. And lately, generative news using tools such as ChatGPT and virtual AI anchors.
The role of AI in news generation
AI will play a major role in the generation, consumption as well as monetization of news. The news media is on the verge of a paradigm shift from human-generated to AI-generated news and we need to face the new reality and adapt to it instead of being wary of it, Kumar explained.
Elaborating, he said the future news workflow could have a scenario where a reporter inputs basic facts about an event in a programmed system (he or she doesn’t write it out and only fills information into a template), where an AI-powered news generator writes out an article. The editor’s job then is to check for narration, facts, or add additional information.
This marks a big shift in the reporter-editor relationship or their roles in news generation and is the future of news. Like it or not, somebody will do it and we all have to follow suit, he explained.
More importantly, the role of humans – which will still remain – in the new order will change from creation to curation, and that will be another big shift. As journalists, we will have to worry about facts, which is more important, rather than the style of writing, which would be taken over by artificial intelligence such as ChatGPT, Kumar said.
We can also think of a scenario where AI can be used to connect different but related news events such as rains, floods, and government advisories, dig into news archives and add value and perspective to a news report. “The possibility of such a scenario is causing upheaval among journalists. But whoever does it first, will move ahead faster,” Kumar said.
Talking about broadcast media, where quite a few channels are trying out AI-generated news anchors, Kumar explained that this will change the way our news studios function because this will eliminate the need for an elaborate set-up that comes at a high cost. And this a natural progression.
AI can be used to add specific such as parameters such as dress and facial features to suit a region, background, or speaking style to create our own preferred avatars of a virtual news anchor, who will read out a readymade script. A host of apps and products are already available for rendering such services, Kumar said.
Personalizing news consumption and ads
In the digital era, news consumption has also become very personalized and this is where AI plays a key role. Unlike earlier, when one had to pay for an entire newspaper irrespective of whether one read just an article or one section, in digital offerings, the whole ballgame has changed.
In the new order, AI can help understand which event or article matters to whom – based on clicks, time spent, comments, likes, and shares – build deep consumer profiles, and help create targeted preferences to deliver the most relevant news and in their preferred format and language to those consumers.
For instance, a choice of formats could include long, or short text only, image only, text plus image, videos, or immersive content. In simple terms, the consumer gets to read and see what he or she prefers – which is kind of already happening on social media platforms such as Facebook to a certain extent.
Ads personalization will move away from cookies, which track our buying preferences, to AI, which can help contextualize ads with an event or a consumer profile – in turn creating more convergence than ever before.
No free lunch
Though AI has the potential to create many positives, there are dangers involved. Kumar pointed to deep fakes, which are worming their way into our lives through news feeds; hallucinating chatbots; and a fear of job losses not just in news media but in other fields where AI is slowly playing an integral part in operations.
Artificial intelligence is a great disruptor across all verticals and not just news, he said, stressing the need for regulated AI for safe disruption. We need to learn lessons from history, where we tend to misuse technology before we use it right – be it nuclear energy or the internet, he said. The same is happening with AI, he said. Facts first, and text later can help deal with some of these issues.
So should we embrace it or fear it? We are fearing what we don’t understand so it is better to understand the pros and cons first, go for meaningful regulations, and then embrace it, Kumar concluded.