Artificial intelligence being adopted in newsrooms around the world

China's state news agency, Xinhua, announced last month that it is integrating Artificial Intelligence into news production to build a new kind of newsroom, in collaboration with the Alibaba Group.

Dubbed “Media Brain”, the platform was developed through a joint venture between Xinhua News Agency and Alibaba Group. It is an open platform where media agencies can apply to share and share their data resources. It brings together cloud computing, the Internet of Things, Big Data and AI technology, deploying them in the process of news production. Proposed applications for the "Media Brain" encompass every stage of news production, starting from finding leads, to news gathering, editing, distribution and analysis of feedback. It intelligently identifies specific scenarios and emergencies such as fires, explosions and traffic collisions, as newsworthy. It will combine data from multiple sources near the scene of the news, such as cameras, sensors, drones etc.

The platform’s capabilities include face verification from videos and images. It can also be used to track copyright violation of all forms of media.

Privacy concerns related to this platform were raised in an article by the managing editor of China Money Network, which itself is an AI-based platform tracking China’s private equity, venture capital and technology sectors.

For a while now, newsrooms around the world have been exploring the use of AI for a range of purposes, from targeted personalised distribution to content generation.

A report from the Reuters Institute for the Study of Journalism at the University of Oxford on Journalism, Media, and Technology Trends and Predictions for 2018, highlights several ways in which newsrooms are adopting AI. One of the most common applications of AI is for improving content recommendations, by understanding individual preferences, and personalising each edition in terms of ‘personalise each edition in terms of format, time, and frequency’. AI driven paywalls can help in identification of likely subscribers and based on previous behaviour serve them an offer which is most likely to persuade them to subscribe. AI could also be used to create more personalised advertisements or push e-commerce recommendations based on previous interest shown.

It also talks about the potential for AI to aid journalists with fact-checking, possibly even in real-time, during a live radio or TV interview.

The use of AI is also being explored for automation of workflows and boosting production efficiency. The report cites a project funded by Google’s Digital News Initiative to generate 30,000 stories per month for local outlets.  It combines editorial expertise and automation, leveraging of growing availability of open data. Once a journalist finds a story using publicly available databases and write a story, multiple versions are then created automatically for different local publications. Graphics, video and pictures for the stories are automatically generated.

The Washington Post has been using Heliograf, its in-house automated storytelling technology to cover Washington, D.C.-area’s weekly high school football games.

In 2014, the Associated Press (AP) entered into a collaboration with Automated Insights, to write earnings reports covering publicly traded companies using algorithms, drawing on information from firms’ press releases, analyst reports and stock performance. A subsequent collaboration between AP and researchers from Stanford University and the University of Washington ‘found compelling evidence that automated articles increase firms’ trading volume and liquidity’.

Chinese tech giant, Tencent, has also been working on using AI for content generation in areas such as financial reporting and sports news, along with understanding user preferences and making recommendations. In late 2015, South China Morning Post reported that Tencent’s automated newswriting program, Dreamwriter, created a business article on inflation, including analysts' comments, in less than a minute. More recently, the executive director of Quartz, Zach Seward delivered a speech at a conference in China organised by Tencent and it was turned into a news story by a combination of AI-based speech to text software, automatic transcription, and Dreamwriter.

Automation is also being used for news detection, as seen from the Reuters News Tracer, which tracks social media and highlights breaking stories to reporters.

Such applications could increase breadth of coverage, help cover local content and allow news organisations and journalists to direct limited resources towards in-depth investigations. AI also has potential as an aid to reporters by enabling rapid analysis of vast troves of data and documents, such as the Panama papers.However, they have also raised concerns that journalists’ jobs could among those expected to be lost to AI.

In this context, a 2017 study from the Reuters Institute for the Study of Journalism on the use of automated journalism in European news agencies found that though most news agencies are already using or actively exploring automation in news generation, only two of the organisations surveyed were using algorithms to compare new information to historical data, and provide interpretations, thereby adding analytical value.

Automation is still not being used widely for more complex reporting. The most popular areas for automation are finance and sport reporting. Currently, there are also limitations with natural language generation ability.

The use of AI for detecting fake news is another important implication of the technology for journalism.  In December 2016, Facebook’s director of AI, Yann LeCun said (subscription required) that AI could be used to detect fake news. The MIT Tech Review wrote about start-up,, which works with publishers, advertisers and networks who do not wish to be associated with false or potentially offensive stories. Last year, a Fake News Challenge was launched through a grassroots effort of over 100 volunteers and 71 teams from academia and industry from around the world, to foster development of tools to help human fact checkers identify hoaxes and deliberate misinformation in news stories. However, it should be noted that this last challenge is still about assisting human fact checkers. The efficacy of AI in filtering out fake news by itself remains to be seen.

In conclusion, AI appears to be here to stay in newsrooms, supporting journalists in their work and not replacing them, at least for the time being. A large proportion of news consumption already happens through algorithm-driven recommendations on social media platforms. Now algorithms are steadily entering into content generation side as well. If used as an aid, it could improve creation and enhance distribution of news. On the other hand, if it encourages owners to further reduce limited human resources, before AI manages to reach the required level for in-depth reporting, or gives governments greater control over what citizens see in the case of state-controlled outlets, AI could have an adverse impact. Then the quality and extent of coverage (or lack of it) will also enter into the list of concerns for public policymakers, as well as civil society. 

Visit site to retreive White Paper:
FB Twitter LinkedIn YouTube