News media is navigating the AI landscape

Even though it was one of the myriad topics of discussion, artificial intelligence (AI) ultimately proved to be the show stopper at the International News Media Association’s (INMA) South Asia News Media Festival in Delhi in July – pointing to the changing technological face of the news media landscape in India, Asia and elsewhere.

From creating and safeguarding content to streamlining newsroom workflow to leveraging technology and AI for resource optimization and revenue growth to building engagement and boosting traffic, the discussions and presentations proved that newsrooms, including those in India, are fast waking up to the AI reality in a big way, and how.

Speakers from India and overseas spoke up for the effectiveness, productivity, and efficiency of AI – in search, revenue, content, and data processing. They explained how they are using AI tools, including generative AI, to maximize operations in whatever way possible and what the future with AI holds for them.

Six applications of AI

Daniel Hulme, chief AI officer, WPP, United Kingdom, in his rapid-fire presentation, explained the six application areas of AI. The first is ‘task automation,’ which includes robotic process automation, chatbots, and object recognition. Then comes ‘content generation’ such as images, video, text, and music; ‘human representation’ – deep fakes, voice, and personas; ‘insight extraction’ – machine learning, data science, and analytics; ‘decision making’ – optimization, decision trees, expert systems; and ‘human representation’ – exoskeletons, avatars, cybernetics. The key challenges and interventions to watch out for are security, safety, governance, and ethics.

“The nation that leads in AI will be the ruler of the world,” Hulme said, explaining how AI tools can be used in the political, environmental, social, technological, legal, and economic contexts. According to Hulme, he does not foresee job losses due to the automation of labor – instead, he sees people moving into less tedious work because many complex and time-consuming tasks can be accomplished through automation.

Smita Salgaonkar, principal architect of marketing technology for Google, felt many companies are contemplating their AI journey but taking that first action step can be daunting. They have to understand their data first and that will determine the success or failure of using GenAI, she said.

News media use cases make workflow more efficient

Sonali Verma, lead of INMA’s Generative AI Initiative, explained how GenAI is changing the news media and shared examples based on chats with about 90 publishers from around the world. For example, Times Internet uses AI to gather first-party data. The group has an advertising chatbot that starts with three interesting questions and follows up with three more interesting questions. It asks for the user’s phone number and tracks the length of the chat, generates FAQs, and gathers other details.

DPG Media uses copy-editing tools that explain why changes were made in the text. Nikkei uses AI to generate summaries and combines multiple articles for readers who may want to know what ultimately happened in a certain developing story.

Sharing data, she explained how 69.6% of her respondents are using AI to generate text, 21.5% for information gathering and sense-making, 20.4% for multimedia operations, 16.6% for business, 8.8% for translation, and 7.7% for data. Other uses include transcription, coding, enhancing user experience and metadata. “Overall, attention is focused on improving and making existing workflows more efficient, with considerably less attention to exploring and innovating new experiences,” she said.

Journalists, she explained, are ready to take AI’s help in analyzing and getting data or information and processing it. However, they do not want AI to meddle in making decisions and solving problems, thinking creatively, and communicating internally with colleagues.

Even if AI is used extensively in media, what will not change is accuracy, facts, surprise, and storytelling; holding power to account, which will remain as important as ever; trusted and stable relationships with individuals and audiences; editorial priorities and real-world interactions such as meeting and talking to people, investigating, and breaking distinctive stories, she said.

News companies should keep in mind the 80/20 rule as they embrace GenAI – allow GenAI to complete 80% of a task but ensure that a human does the critical 20% – “You want a human in the loop and the beginning of the task and at the end of it,” Verma said. This ensures the task begins with the correct prompts and direction and also checks for accuracy when it is completed.

Resource optimization, revenue and reader growth

In another session on leveraging technology and AI for resource optimization and revenue growth, Puneet Gupt, COO of Times Internet, revealed the impact of the AI-led world on news media and how Times Internet is navigating this disruptive change. “When people start to talk about AI, the first thing they usually talk about is loss of jobs,” Gupt said, adding no jobs will be lost if we use AI to create revenue.

Gupt identified falling eCPMs, losing engagement to ‘doom-scrolling,’ building first-party data, combatting the fake news narrative, balancing ads and subscriptions, team productivity, stagnating traffic, and overcoming the technology barrier as some of the key challenges in the news media industry. He went on to explain how AI can be effectively utilized to overcome many of these conditions.

From building a chatbot that improves results for advertisers as mentioned earlier, to using AI-generated creative ads that showed a 4x increase in click-through rates, to providing accurate gender prediction, Times Internet has seen dramatic results. These have had a direct and positive impact on economic returns.

The company has seen a 95% lift in CTR on personalized push notifications and has seen recirculation of older stories increase. In the newsroom, GenAI has increased productivity and cost efficiency, completing tasks such as generating first drafts for editors to review and converting articles to video, Gupt said. He particularly mentioned an ad campaign created with AI that ran on billboards.

Prashant Verma, partnerships director, Dow Jones, said the company, which has leading brands such as The Wall Street Journal and Dow Jones Newswires in its kitty, has been using AI to serve the specific information needs of consumers. It uses AI for metadata applications such as tagging for more than 600k articles every day across 3,000 subjects and 200k+ company codes apart from highlighting trending news, auto-translation of content, and sending personalized email digests to readers.

Google AI Overview

Rudra Prasad Kasturi, chief strategy officer/digital business for TV Today Network, in his presentation, explained the working of Google AI Overview (AIO). The relatively new tool sources information from various websites with links to the sources, and includes the option of asking a follow-up question to improve the search. The search goes beyond text to include images and video.

How does Google AIO work? When one types in a search query, such as a question or a request for information, Google uses advanced language models to understand the meaning and intent behind your query. The system finds relevant documents, web pages, articles, images, and videos related to the query. Google also looks at content from related searches and information that others found useful for similar queries.

An AI model then combines all this information to create a response that is directly helpful to the person asking the question. The AI adjusts its responses based on the specific context of your query, providing tailored summaries tempered from what Google learns from earlier user interactions with search results. If users prefer certain types of responses, the AI adapts future results to match these preferences. The AI continuously updates its understanding with new information and user interactions, improving the accuracy and relevance of its responses over time.

However, one should note that the AI Overview is not searching the internet in real time for documents. Instead, it relies on pre-existing training data, and Google’s index and knowledge graph to provide information.

These are just glimpses of what transpired at the two-day Inma conference at the Le Meriden. We have earlier written that AI is a technological evolution that the media is compelled to investigate, adapt, and adopt. From what we heard and saw in the two-day event, the media has moved quickly beyond a mere exploration of tools to a level of experimentation, immersion, and implementation. As Gupt said, we are rapidly moving from ‘Save the media from AI’ to ‘Save media through AI.’