AI for Journalists: Tools, Ethics, and the Future

Last month, Google pitched an AI tool for journalists that would be able to generate news articles for various news outlets, including The New York Times, The Washington Post, and News Corp, which owns The Wall Street Journal. According to Google, the tool would not replace journalists in reporting or creating articles. Still, it would serve as an assistant to journalists, allowing them to use technology to improve their productivity and efficiency. Google has also said in a statement that it is looking to partner with news outlets focusing on smaller outlets for the tool, which they internally refer to as “Genesis.”

This news comes as part of a growing movement of journalists and news organizations incorporating AI into their work processes. With the recent popularity of generative AI models and the use of AI to streamline various industries, it is unsurprising that technology has finally caught up with journalism. News outlets like Bloomberg, AP, and The Washington Post already use algorithms and AI to dissect reports and data and produce articles on these reports. With the introduction of AI into newsrooms, it is becoming essential to learn how to use AI responsibly and effectively without sacrificing journalistic principles.

The best way to use AI as a journalist is as an assistant who can make the more mundane tasks go easier and faster, leaving journalists to focus on the actual writing process. There are many tools out right now that can do this. ChatGPT, one of the most popular tools today, can serve as a sounding board for story ideas and pitches and provide a way to play around with audience and tone. The assistance of ChatGPT and other generative AI tools make it especially useful during periods of writer’s block since it makes it easier to streamline brainstorming and other processes associated with writer’s block.

Several AI tools today also automate other repetitive tasks in writing. Tools like Descript simplify the transcription of interviews and other documents needed to write a story. Others like Jasper AI help to make the writing process more efficient by generating social media content, emails, and SEO-optimized articles. Jasper AI also has a chatbot that assists with tasks and does research quickly.

AI is especially useful when working with data, especially in data journalism, to break down and analyze large amounts of data. Many newsrooms already use AI to tell visual stories and generate news articles from data. One project by two investigative journalists at the Norwegian Broadcasting Corporation used AI to investigate how TikTok’s algorithm shows or hides videos about the war against Ukraine depending on a user’s location. 

Newsrooms across the US have progressively gotten smaller over the last 20 years as the media landscape has grappled with the consequences of online publishing and the decline of print advertisers. According to a Pew Research survey, newsrooms shrank by 23% between 2008 and 2017, with local publications suffering the brunt of the employment decline. In this area, AI has the potential to help revive local newsrooms by allowing publications and reporters with limited resources to streamline their work process and channel their energy toward the work that really matters to them. Organizations like Partnership on AI (PAI) are helping do just that by compiling AI tools for local newsrooms and developing best practices for the use of AI in newsrooms.

AI is also helping to simplify other aspects of the publishing process, including the unpleasant parts. This includes moderating and filtering comments to eliminate harassment and encourage productive conversations. Tools like Conversario help with community management and filtering out spam, allowing reporters and editors a safe space to share the news. Other tools such as The Washington Post’s Heliograph and Reuter News’ News Tracer are also helping journalists to spot and validate news and analyze data trends.

The best way to use artificial intelligence in journalism is as a means to get unwanted and repetitive tasks out of the way and focus on things that we want to get done, like actual reporting and writing. Here are some tools to help maximize productivity and make writing easier.

• Quillbot - an AI-powered tool that paraphrases what you say. Quillbot is also able to scan texts and correct errors in grammar, spelling, punctuation, and word misuse.
• Murf.ai - a versatile AI voice generator that can turn articles and text into audio for podcasts.
• Connexum - an AI-powered news engine that allows journalists to source real-time multilingual headlines, articles, and summaries extracted from thousands of news outlets.
• Pinpoint by Google - a research tool for exploring and analyzing thousands of documents and transcribing audio.
• Postwise AI - an AI-powered social media tool for crafting and scheduling engaging posts.
• JECT.AI - funded by the Google News Initiative, JECT.AI generates ideas, angles, and content from an archive of millions of news stories and articles.
• Vetted  - an AI tool that provides access to verified sources.
• Narrativa - a generative AI tool that generates written content, including outlines and article drafts. 

As with every other aspect of journalism, using Al requires some considerations as to how to use these tools responsibly and ethically. One primary concern with generative AI is its tendency to hallucinate and generate biased content, which may impact the accuracy and credibility of the finished article. This means that these tools should be approached with caution, especially when developing content for articles and news stories. There is a responsibility to an audience to crosscheck and fact-check when using generative AI tools. In these cases, a healthy amount of skepticism is necessary.

When using these tools, it is also important to consider who is accountable for AI-generated content. With content written by a human, it is clear who should be held responsible for errors and missteps, but this gets murky with AI-generated content. And without being able to identify who is accountable for a fumble or harm, it gets harder to correct and improve the situation. An ideal scenario would be a perfect balance of AI-generated content with human editorial control to check and correct the  AI-generated content. Transparency with AI-generated content also goes a long way in solidifying trust, which is essential in the relationship between a journalist and their audience.

The constant hype of AI as a technology capable of doing any intellectual human task and eventually replacing humans in most jobs can sometimes be a distraction from the actual capabilities of AI and the ways that we can utilize it to work for us. Like any other technology, AI can be used in a way that is beneficial to both writers and our audience.  This might mean engaging with it in an ethical way, like we would with any other aspect of journalism. If we use it collaboratively or as just another tool to make writing easier, we can then make it work for us.