Role of AI and Machine Learning in Journalism

Digitisation and artificial intelligence together, in recent times, has changed the face of most industries. One industry, in particular, has seen monumental changes in journalism. The realm of journalism is enormous, and the art of covering important social events is very difficult and challenging. Also, it is a known fact that humans are the ones that carry out the majority of this work; machines, in most cases, care less. Cut to the current scenario, strangely enough, a large portion of the journalistic industry depends on technology. When we use the term technology here, we refer to the superheroes’ artificial intelligence and machine learning.

Technology is essential in keeping the news and journalism industries buzzing, considering the pace at which news is produced on a daily basis. From communicating with news sources in real-time and content generating first-hand in image, video, and text form to publishing the content across an ever-increasing array of channels, the role of AI and machine learning in journalism has come a long way.

Recently, the process of generating, producing, publishing, and sharing news has transformed dramatically due to the increased use of AI by news companies. And thanks to the cutting-edge innovations in artificial intelligence, we have already reached a phase where machines are producing news items, pretty much like the one you are reading.

Before we discuss the impact artificial intelligence and machine learning have had on journalism, let’s rewind to a time when it all started and which publication made the most of machines.

Quakebot – the AI bot that Reported an Earthquake within 3 Minutes

Back in 2014, the Los Angeles Times made a massive impact in the breaking news segment when it reported an earthquake just after 3 minutes. If you are wondering about the reporter who pulled it off, well, it wasn’t exactly a human who accomplished this feat.

The report was published by an AI bot called Quakebot which was designed by the same publication with a vision to report an earthquake immediately. The software was designed in a manner which enabled it to examine earthquakes issued by the U.S. Geological Service. After reviewing it and when specific requirements are met, a draft is automatically generated. When a Times editor decides the post is noteworthy, the report is released after informing the staff. And guess what? The Quakebot actually has an author page under which the articles are published.

Recommended: UNC Journalism School Standardizes on Nureva for Classroom Audio

Fast forward to the current times, each week, mainstream media sites publish hundreds of thousands of stories that have been written by AI. Several renowned media organisations today have created internal versions of the Quakebot. The Washington Post uses Heliograf, the BBC uses Juicer, and roughly one-third of the information published by Bloomberg is produced using a system called Cyborg.

These programs use data in the form of graphs, tables, and spreadsheets. They examine these to glean specific facts that can serve as the framework for a story. After creating a roadmap for the article, they begin generating sentences using natural language generation software.

These algorithms are only capable of producing articles that include highly structured data like footage of a football game, or spreadsheet data from an annual report of a firm. They are unable to produce articles with originality, creativity, or in-depth analysis. Unlike what most people feared, AI has not eliminated journalists from the system, instead, they have dramatically increased the production of specialist papers.

Kenn Cukier, a senior editor at The Economist, summed it up quite nicely. He says, “We can’t be precious about this: it’s about what is best for the public, not what is best for journalists. We didn’t cling to the quill in the age of the typewriter, so we shouldn’t resist this either. It’s a scale play serving niche markets that wouldn’t be cost-effective to reach otherwise.”

A report by Grand View Research predicted that the market for AI in media and entertainment was valued at USD 10.87 billion in 2021, and from 2022 to 2030, it is anticipated to expand at a compound annual growth rate (CAGR) of 26.9%.

The primary driver of the market is the growing acceptance of virtual creation in the media and entertainment industries, as well as its capacity to produce real-time virtual worlds and high-definition images.

The Impact of AI and Machine Learning in Journalism

The most noticeable feature of this phenomenon is news automation, which has surely sparked the hottest discussions among journalists. The concept of “robot journalism,” as it is frequently referred to, is looked upon as both progressive as well as regressive at the same time. We’ll reserve the discussion for this part later as this article aims to bring forth the positive impact machines have had on journalism.

Read: 5 Biggest AI Tools You Should Be Using Right Now with ChatGPT

In automated journalism, they may compile and distribute the material at the touch of a button in addition to gathering information and comprehending data pools. Stories are produced at scale by algorithms.

The rise of artificial intelligence in the media and entertainment industry is fueled by applications in gaming, fake story detection, plagiarism detection, production planning and management, personalization, sales and marketing, and other sectors.

AI-generated Articles

With more difficult tasks like detailed analyses, lengthy pieces, and investigative journalism, AI can assist human reporters. Nowadays, machine-written articles can only be found on straightforward, formulaic subjects like stock market outcomes, sports game scores, etc.

Several renowned media organisations today have created internal versions of the Quakebot. The Washington Post uses Heliograf, the BBC uses Juicer, and roughly one-third of the information published by Bloomberg is produced using a system called Cyborg.

These programs use data in the form of graphs, tables, and spreadsheets. They examine these to glean specific facts that can serve as the framework for a story. After creating a roadmap for the article, they begin generating sentences using natural language generation software.

The Washington Post began producing 300 brief stories and alerts about the Rio Olympics using its in-house artificial intelligence software, Heliograf. The Post used Heliograf to create about 850 articles in its first year. These include 500 election-related items that received more than 500,000 clicks. While this may not seem like much in the grand scheme of things, the majority of them were stories the Post wasn’t going to allocate employees to in the first place.

The Juicer not just takes news content but also automatically tags it before offering a full-fledged API to access the data and content. BBC describes it as a news aggregation and content extraction API.

Bloomberg Cyborg is a natural language processing technology known as Bloomberg Cyborg. It is intended to assist users with obtaining knowledge and insights from significant amounts of text-based data, including news stories, research reports, and corporate documents. It can be used to keep tabs on news and market trends, monitor business and industry performance, and spot prospective investment possibilities.

These algorithms are only capable of producing articles that include highly structured data like footage of a football game, or spreadsheet data from an annual report of a firm. They are unable to produce articles with originality, creativity, or in-depth analysis. Unlike what most people feared, AI has not eliminated journalists from the system, instead, they have dramatically increased the production of detailed, niche articles.

Transcribing Audio and Video Interviews

Software that intelligently transcribes audio and video recordings into text is one of the most practical applications of artificial intelligence (AI) and machine learning (ML). Natural language processing (NLP), a subfield of artificial intelligence that focuses on the study and use of methods and tools that let computers process, analyse, understand, and reason about human language, is the foundation upon which AI transcription software and services are built.

AI can help journalists with the time-consuming task of transcribing audio and video interviews. The Ai-powered software can turn auditory data into text so that journalists can concentrate on drawing conclusions.

Below are a few examples:

SpeechText.AI: It is an example of software for interview transcription with near-human accuracy. The software is compatible with all formats of audio and video files. Besides transcribing, it can also help with transcribing live meetings and interviews.

Sonix: The software can transcribe your interviews in 38+ languages using our state-of-the-art transcription software. The automated system takes care of the rest after you upload your audio or video files. Sonix supports several formats like mpg, avi, wav, mov, and mp3. In just a few minutes, you will receive an email along with a link to your transcript.

Otter.ai: The AI-powered helps you with full audio transcripts, highlights, and automated summaries. It has the potential to translate hours of audio and video recordings to text in minutes.

Speak AI: Speak is an ideal choice for an AI transcription service as it gives you numerous options for gathering significant audio or video data and converting it into insights, sans any code. Speak allows you to simply upload locally saved files and create custom audio and video recorders that can be embedded.

Beey: It can convert podcasts, videos, online meetings, interviews, recorded lectures, meeting minutes and other online files. Modern captioning technology makes it simple to create captions and subtitles of high quality. The platform’s reach is genuinely global because it supports more than 20 different languages.

Recommended: 35 Generative AI Tools for 2023 That You Should be Using Right Now

Flagging Alerts

Another example of how artificial intelligence and humans might collaborate to produce better outcomes is one of the most recent projects in this area. When a trend or anomaly in a huge database is discovered using AI, warnings can be sent to journalists. It can offer publishers and content creators tools to spot fake news and decrease its impact on their readership.

Thanks to the efforts of start-ups like Logically AI is now being used to stop the spread of false information on the internet and social media platforms. Founded in 2017 by Lyric Jain, the UK-based start-up aims to create a solution to validate the accuracy of the news, online conversations, and visuals. Also, the business offers a Chrome browser extension for fact-checking news articles on more than 160,000 social media and news websites. To comprehend and interpret text, Logically’s AI algorithms use natural language processing.

Controlling Bias

The media cannot avoid bias because it is a worldwide problem. Yet, because AI’s machine learning algorithms are trained to take accuracy into account, it helps to lessen the subjective interpretation of data by humans. The Financial Times also uses AI to identify patterns in markets and in individual economies. In order to lessen the bias in its own output, The FT also makes use of bots.

The Bipartisan Press claims to have developed an AI that can detect whether biasedness in text with an accuracy of more than 96%. While explaining this, Winston Wang, managing editor at the Bipartisan Press, stated that for ease of use, we adopted a scale of -1 to 1, where a negative value indicates left bias and a positive value indicates right bias. The degree of bias is indicated by the result’s absolute value. For instance, a piece with a 0.8 has a substantial pro-right bias.

The team is presently developing a Chrome browser plugin, according to Wang. This might be utilised to inform readers about the biases contained in the news stories they read. Better content recommendation systems might theoretically be implemented using artificial intelligence.

Bots in the Newsrooms

Robots have not only entered the digital marketing sectors and sales teams but also newsrooms. China’s state-run Xinhua News Agency had a breakthrough when it deployed a robot as an English AI Anchor. The newest anchor Ren Xiaorong is believed to be endowed with the professional abilities of a “thousand presenters.”

She asserts that she is capable of providing “news broadcasts about any topic all year long.” She further adds that without stopping, she’ll broadcast news nonstop for a full year—365 days and 24 hours—without taking any breaks. I will be broadcasting news for the entire year, round-the-clock, without rest, for 365 days, 24 hours,” Xiarong continues. You can always find her on news websites or in the studio. Every discussion and piece of advice you provide me will only advance her knowledge, she adds.

AI as Click Bait Detectors

Clickbait is one of the most common challenges in journalism today. To put it in simple terms, the term clickbait can be defined as an article headline that uses sensationalist language to persuade readers to click over to a particular website or webpage. The website then cashes on the user’s activity data or makes advertising money off of their clicks.

In the hyper-aggressive digital environment, our social media accounts, apps, and even emails are swamped with all kinds of news and content. And each of these articles, content is vying for our attention. Because it’s so simple to share and repost content on social media, clickbait has run amok.

To accurately distinguish between clickbait and non-clickbait headlines, machine learning methods like Naive Bayes, Logistic regression, and SVM are used.

Final Thoughts

With each new technological development, journalism keeps evolving. Mass printing replaced handwriting. The telegraph accelerated the gathering of news over long distances. The telephone and radio propelled journalism even more.

Also, the news industry has altered as journalism has transitioned from radio to television to cable to the internet in the last 100 years alone. In this latest version of the tech revolution where we are witnessing the rise of stalwarts like ChatGPT 4, Google Bard, etc, the future of journalism looks soaring high.