Harnessing the might of AI for a more efficient newsroom

AI in newsrooms

AI has numerous applications across newsrooms and this form of cyborg journalism is here to stay. Roughly a third of the content published by Bloomberg News is generated using a system called Cyborg. According to Bloomberg’s Editor-in-Chief John Micklethwait, there are 4 areas where AI is significantly changing journalism - automation of repeatable events, techniques of the trade, signalling (machines scour social media for trending issues) and personalization (providing readers with news of their preference)

1. Automation in repeatable events

AI can write news stories that are based on highly structured data. Raw data about news events can be transferred into pre-programmed templates to form a coherent story. Automation is especially useful in beats like Weather, Finance and Sports, where figures play a pivotal role. AI bots can scrape this data as it is released and the information is then matched to relevant phrases in a story template. During the Rio Olympic Games 2016, the Post’s Heliograf put together the news story by analyzing data about the games as it emerged. The resulting stories had a similar core structure but varying details.

2. Techniques of the trade

AI helps journalists to focus on their area of expertise by taking over monotonous and laborious tasks such as transcription of recordings with the help of voice recognition technology. AI can help journalists analyze large data sets quickly and without any errors. It can look for patterns, crunch numbers and provide journalists with vital results and background information. Reuters’ Lynx Insight can alert a journalist that Walmart shares fell by ten per cent, noting the last time they dropped as fast and why, by pulling in headlines from that day about the company.

Automatic translation has been a boon to foreign correspondents who typically report in one language and write in another. With superfast machine translation,

AI is the driving force behind numerous publishing and content marketing tools. Audience engagement tools utilize AI to help publishers handle spammy or toxic comments. One such tool, Metype has integrated with Jigsaw’s Perspective API to provide moderators with a toxicity score based on which publishers can easily moderate comments on their posts. Metype has also integrated AI to automatically tackle spam in the comments with minimum human intervention.

Publishing tools like WRU.ai use AI to optimize editorial performance by providing precise data-driven predictions.

Functionalities like tagging articles with all the relevant tags and churning out data for which articles need more marketing spend are hence automated. Frase is a tool that helps publishers create better content, faster by allowing them to focus on the topics that matter to their audience. It leverages AI, machine learning and NLP to help publishers discover topic gaps in their existing content and also to generate summaries of research topics.

3. Signalling

Since news globally has been breaking primarily on social media, AI can effectively keep an eye out for trending events, issues, disasters, shootings, etc. on these channels. Reuters’ News Tracer combs social media posts to flag potential breaking news, and rates its newsworthiness. It also verifies the news like a journalist would, by looking at the poster’s identity, checking whether there are links and images, besides taking into consideration other factors.

Alerts can be set for any anomalous information or patterns that bots come across. Journalists can then take a call on whether the story is newsworthy enough to be pursued. During the Olympics, for instance, The Post set up alerts to inform editors if a result was 10 percent above or below an Olympic world record.

4. Personalization

Due to its ability to personalize content, the use of AI not only helps with reach but also with audience engagement. “It’s useful to find out how well your stocks are doing, what the weather is like in the city where you live and how well your soccer team is doing.” says Micklethwait. Chatbots after continued user interaction, can recommend news and stories that are personalized to meet a user’s interests. The BBC used such bots to help cover the EU referendum and the Oscars. “What’s noticeable is that when people engage with in-story bots, they usually ask up to five or six questions.” said Paul Sargeant, BBC’s news visual journalist.

AI systems monitor users' likes, dislikes and reading patterns. After learning readers’ preferences content can be tailored to better suit their interests and to push them towards related content. Users appreciate and encourage personalization as it helps them cope with information explosion.

AI allows consumers to control what content reaches them, when it reaches them and in what format it reaches them. By subscribing to AI powered RSS feed readers, consumers can get all the updated content from their favorite sites in a manageable form at one place. With 14 million users, Feedly is the largest RSS reader on the market. Users can specify what RSS feeds they would like to follow and Feedly will collect every new headline those sites churn out, offering anything from a snippet of information to the full story, depending on how much the publisher allows.

Personalization is also key to increasing subscription revenues. It also appeals to publishers because the technology used to target content to readers can also be used to target advertising. This will lead to an increase in revenue for both advertisers and publishers. Since AI can process consumer behavior patterns with precision, it can help create individual-centric campaigns. Businesses will get better returns as their ad campaigns, which are now more streamlined and personalized will receive maximum engagement.

A new hope against fake news

There is a crisis of trust in media and it is impacting readers’ willingness to pay for news. Harnessing AI’s power to assure data accuracy and to identify fake news is the only way to turn this situation around. Machine learning and NLP with their potential to learn behaviors through pattern recognition can assess content and sources for credibility and objectivity. They can help readers identify partisan content or hate speech, which is often the product of a potentially fake news source. AI can send alerts to fact-checkers when certain articles or sources they scour are flagged as inaccurate by a substantial number of people. If these human reviewers find an article or source to be fake, they can then exercise judgement and alert readers to the same. The Quint’s Webqoof, a section dedicated to fact-check online information and debunk fake news, is extensively powered by AI.

It’s alive! Will AI take away your jobs?

According to Tom Van de Weghe, a research fellow at Stanford University, AI in its most basic form is a system that makes autonomous decisions, performing tasks that mimic acts of human intelligence like solving problems, understanding language or recognizing sounds and objects.

To help computers perform their specific tasks, they need a set of step-by-step instructions that tell them what to do: an algorithm. These algorithms are set up by humans. AI could become the savior of the trade — making it possible for journalists to better cover the increasingly complex, globalized and information-rich world we live in.

The important thing to remember here is that this process cannot happen without a human journalist present who, with a goal in mind, asks relevant questions about the data. So, NO, AI is not going to take away your jobs. However, reporters and editors need to swiftly learn how to integrate these systems into their workflow in order to create what Reg Chua, executive editor, editorial operations, data and innovation at Reuters calls “cybernetic journalism”.