Artificial intelligence (AI) has grown by leaps and bounds in the last decade. The use of AI has permeated almost every sector, and boardrooms often see animated discussions on how AI can be leveraged into their business. 

However, not all industries have advanced AI technologies streamlining their processes. Some industries are beginning to dip their toes into integrating AI into current systems, while some others have detailed strategies on AI adoption. 

According to a recent survey by Journalism AI, more than 71 news organizations in 32 countries are currently incorporating AI-related technologies into their daily operations. The Washington Post’s in-house automated AI technology – Heliograf – has churned out more than 850 articles in a single year. These range from reports about the Olympics to covering football games and elections. 

AI-assisted journalism : Next frontier of discovering stories

News broadcasters have seen varying degrees of AI adoption over this time. Since newsrooms play a pivotal part in our daily lives, it is only natural that their industry leaders think of using AI in their regular operations to streamline their processes.

Incorporating AI methodologies into journalism will automate effort-intensive tasks like secondary research, transcribing, and data entry so that journalists focus their attention on what they do best – sharing stories.

Newsrooms can utilize AI to analyze massive amounts of data and pick up patterns that would take hundreds of hours if done manually. This is particularly important to journalists who need to go through large volumes of information and archived research pieces of information to fine-tune a story.

Where is AI currently being used in newsrooms?

Three key areas where AI is being used in newsrooms are newsgathering, news production, and news distribution.


News broadcasters worldwide use AI-enabled technologies to sift through vast reams of data from various sources to gather content relevant to the current context. AI tools can pick up fascinating stories and trend connections that may not be immediately obvious to humans. An example where investigative journalism relied on AI comes from Mexico, where homicide news appearing on the web was collated to match with homicides reported. This was a significant breakthrough in understanding crime rates in the country.

The aspect of investigative journalism where reporters get essential information from a stockpile of data and make sense of it has its charm. But in reality, this is akin to finding a needle in a haystack.

For AI to detect something as a useful piece of information, it needs to have rich metadata content that supports to parse through the large volumes of material. Hypertagging engines that use active learning can help in auto-tagging this piece of information for further use. Other applications that invite the use of AI include sourcing information, monitoring a specific issue, fact-checking, or finding relevant information around a news piece.

For example, the New York Times introduced an AI tagging and annotation tool called Editor to streamline the editorial process, by suggesting that one look through specific archives find information relevant to a current story.

News production

AI can help in automating repetitive and similar tasks in the production process through image recognition, metadata tagging, speech-to-text conversion and editorial automation.

Through active learning and utilizing the editorial metadata generated, AI can help news creation teams with custom-made editorial workflows.

For example, Voitto, the Finnish robo-journalist, uses machine learning algorithms to produce textual and illustrative content that is carefully curated to fit various newsletters and social media accounts.

News distribution

Engagement data, when paired with AI-extracted content parameters, can be analyzed to gather insights that can help news broadcasters produce content that would be more popular among viewers, and would boost engagement on different distribution mediums.

For example – Anchor duration/ story genre could be analyzed together with TRP ratings of the news shows to find out which aspects of the content are actually making it work.

AI tools can also be used to debunk fake news by extracting information, retrieving images, and conducting image forensics. For example, The Times employed an AI-powered digital butler called James to distribute personalized news and increase audience engagement. For many months, the tool harvested large volumes of data to analyze user habits and personalize content. 

AI can also help news broadcasters in developing contextual advertising capabilities and ensuring brand safety of advertisers. AI-enabled analytics can help extract content information around the news which would match with the brand message.

Possible Pitfalls: What Newsrooms should know when practising AI

Though experts around the world unanimously agree that AI technologies would help rethink how news is created, news companies must also be aware of the underlying ethical concerns with AI-related journalistic practices.
As AI tools become increasingly ubiquitous across every aspect of news, it has become essential for industry leaders to be aware of instances where it can backfire. Here are some examples that one should keep in mind:

Filter bubbles

AI-based algorithms can curate the newsfeed to show viewers only what they want to see. News items that are highly engaging to the user are given more prominence, which filters out a lot of content from the users. On the face of it, this depth of content curation might come across as a useful feature, but it is in fact responsible for creating an ideologically segregated group of people that have extreme biases and decreased tolerance to views that differ from their own. 

Google has been in the news for accidentally including filter bubbles into their news results. A search engine called DuckDuckGo asked volunteers to search for a particular phrase at the exact same time from their regular accounts, and then in incognito mode after logging out of their accounts. It was observed that most users received inconsistent results, and being in incognito mode did not affect their results. 


Using AI and customer data to personalize news and curate content to hyper-target the customer’s experience based on previous patterns, demographic, and behavioural data will surely increase viewership of content. It also limits viewer exposure to different types of content. Targeting a specific set of people in the news or having too many news articles for one special keyword can make the story look super-tailored and unauthentic.

Personalizing a news feed is different from personalization in case of entertainment. Ethics of journalism come into effect, where it is crucial to showcase news that matters, rather than news which only engages.

Algorithmic bias

Algorithms used to curate content are ultimately crafted by humans and may reflect their biases in the form of their output. To put things in perspective, the biases in training data can reflect some results that might put news companies in an awkward situation. Google Vision, which labelled a particular thermometer differently based on the skin tone of the person holding it, is a current example where such biases crawl in and are not something that should be taken casually.

Deep fakes and fake news

Deep fakes are AI-created fake audio or video recordings that look and sound genuine enough. They may be crafted to increase viewership, or to damage reputations, or with the intent of misleading viewers deliberately. These are often created using AI and deep learning techniques. 

While AI certainly helps streamline tasks for journalists, it can also create false information quickly. As accessibility to technology becomes more natural, the possibility of misuse rises. Newsrooms should equip themselves to fact check such pieces of information.


Industry leaders in journalism are looking forward to leveraging the potential of AI and machine learning in newsrooms. They wish to go beyond automating routine tasks and ultimately train the machine model to create high-level content for media outlets using AI.

With that said, while AI tools can certainly augment journalism, but they can never replace the age-old process of crafting eye-grabbing news pieces.

While AI technologies help break down the silos present between different departments in newsrooms, an awareness of the ethical and economic pitfalls is most crucial. Good subject expertise can go a long way in ensuring hat the pitfalls caused by integrating AI do not overshadow its many data-driven benefits.