How AI Changes Media Landscape and News Delivery

Despite fear-mongering around the artificial intelligence (AI) technologies and how they’ll affect our daily life and future employment, the real-life AI use cases prove otherwise: AI brings us more benefits than dangers. From saving time and money with automation to making faster and better business decisions thanks to cognitive technologies to making data-based predictions to providing highly personalized user experiences–AI proves to be able to improve any company’s top line growth and increase bottom line savings. And, what’s even more important, AI empowers digital transformation that many businesses are seeking to achieve today to gain an innovation advantage in the future.

Let’s see how exactly AI changes modern media and news delivery landscapes and what it can do better than humans.

AI can write quality news articles

The Guardian Australia uses AI to write news articles. The principle of the algorithm is similar to that used in OpenAI: it is capable of processing large amounts of data, as well as finding and creating text patterns.

In February 2019, The Guardian published its very first AI-written article titled “Political donations plunge to $16.7m — down from average $25m a year”. AI ​​analyzed data on the volume of donations attracted by all political parties in Australia (provided by the national election commission). Then it briefly described its analysis findings, compiled a rating of parties based on the donations raised, added links to the original sources and even scheduled the article publication.

Image source: The Guardian

In fact, if we look at the article, we’ll never guess it was written by AI, not by a human journalist. So yes, AI can write news stories and they can be as good as the ones written by humans.

AI can be a good news anchor (not very emotional, though)

The first AI-based TV presenter appeared on Chinese television at the beginning of 2019. Looking exactly like Qu Meng, a female employee of the Chinese state agency Xinhua (who was used as an AI prototype), the virtual anchor named Xin Xiaomeng debuted in TV news on February 20, 2019.

Thanks to machine learning, Xin Xiaomeng is able to read texts from the screen, as well as learn from living colleagues, imitate their facial expressions and tone of voice.

Earlier at Xinhua, a TV presenter named Qiu Hao had already been invented; his appearance was copied from a real TV anchor Zhang Zhao.

Both AI anchors were developed by Xinhua jointly with the Chinese search engine Sogou. Since November last year, Qiu Hao has been integrated with the agency’s official website, its online TV platform, mobile applications and WeChat public account.

The only drawback of such AI-based news presenters is that they only move parts of the face and aren’t yet capable of gesticulations.

AI can help monetize content with automated media recommendations

One of our company’s clients is a leading UK-based media agency that hired us to build an AI-based platform similar to 9gag and Reddit alongside a mobile application. The platform uses data analytics and machine learning to analyze individual user’s preferences and to help them discover new content (the most relevant, interesting, and eye-catching) accordingly.

In the first six months after the deployment, click-and-go for media partner offers grew 20% compared to the previous period, while the overall audience increased by approximately 40%.

AI can detect and filter out fake news and stories

AI is powerful enough to prevent fake news dissemination. Since 2015, Thomson Reuters has been using AI-based News Tracer to separate real news from junk, unsolicited/disguised ads, and information noise. Now the algorithm sifts more than 700 million tweets per day.

In the process of searching, the AI ​​“thinks” like a person. First of all, it checks the profile of the user who tweeted the news. The algorithm checks whether the user account is verified, who subscribes to it and who the user follows, whether there are any links and images in the tweet, and so on.

Over the past years, News Tracer has helped many Thomson Reuters journalists all over the world. In April 2016, when 77 people died as a result of the earthquake in Ecuador, the algorithm gave the Reuters reporters 18 extra minutes to gather as much information as they could before other news agencies learned about this disaster and rushed to highlight it in their news reports.

AI can report the company’s profit and loss

Data science is a classical AI use case: for instance, back in 2014, the Associated Press (AP) agency found a way to automate the most dreary and boring job for many journalists — analyzing companies’ quarterly reports.

Yelp earnings report created by AI

AI-based software writes technical news about the financial results instead of journalists. Thanks to it, AP managed to increase the number of such news from 300 to 4,400 in just one year. Apparently, AI ​​was able to analyze exactly ten times more quarterly reports than journalists did (3,000 instead of 300).

The introduction of the AI-driven bots also helped AP reduce the number of typos and factual errors, and freed up to 20% of journalists’ time for less routine and more creative work (this did not lead to the dismissal of former employees). In 2018, Bloomberg and Reuters started using similar AI systems.

As we can see from the above examples, when used for good purposes, AI can be a great way for media businesses to save millions of dollars with automation, machine/deep learning algorithms and predictive analytics, and, what’s even more important, it can release the burden of routine and repetitive tasks off the journalists’ shoulders and free up their time for creative work and better user engagement.

And what’s your take on this? Are you for or against the mass adoption of the artificial intelligence tech in media?