Revolutionizing Journalism: The Power of Generative AI in Newsrooms

Table of Contents

1. Introduction
2. Practical Applications of Generative AI in Newsrooms
   1. Productivity and Efficiency
   2. Content Summarization
   3. Editorial Check on Biases and Gaps
   4. Providing a General Sense of Why the Public Might Care
   5. Retaining Human Oversight and Input
   6. Newsletters
   7. Tools for Headline Creation and SEO
   8. Reformatting Text
   9. Scraping Newsroom Archives
3. The Future of Generative AI in Newsrooms
4. Conclusion

Practical Applications of Generative AI in Newsrooms

As the world becomes increasingly digitized, newsrooms are turning to generative AI to help streamline their processes and workflows. While there is ongoing debate about the ethical and transparent use of these tools, many newsroom leaders have found practical applications for generative AI that center around productivity and efficiency.

Productivity and Efficiency

Newsroom employees and journalists are often overworked and burnt out. Generative AI offers several opportunities for decreasing the cognitive workload on individual journalists or newsroom employees, allowing them to spend more time and energy focusing on serving the needs of their respective communities.

One way that generative AI is being used to increase productivity and efficiency is through content summarization. Newsrooms are using these tools to provide lists of key takeaways from lengthy documents, allowing journalists to quickly and easily digest important information.

Editorial Check on Biases and Gaps

Generative AI is also being used to provide an editorial check on biases and gaps in reporting. By analyzing text for common topical features and themes, these tools can help newsrooms identify shortcomings in their reporting, sourcing, and source diversity. They can also provide a general sense of why the public might care about the issues that are discussed and covered in their journalism.

Providing a General Sense of Why the Public Might Care

Oftentimes, journalists and editors can lose sight of the bigger picture. While they have their head buried in their work, the value is self-evident to them. However, it's sometimes not as clear to members of the public why this is such an important issue. Generative AI can help newsrooms provide a general sense of why the public might care about the issues they are covering.

Retaining Human Oversight and Input

Despite the many benefits of generative AI, it's important to retain robust human oversight and input across all of these processes. One example of this is The Marshall Project, which has been using chat GPT to turn complex bureaucratic text into simple summaries as part of the public service journalism that they produce. They use the tool to do textual analysis and help classify common topical features and themes in a particular body of text. They also give it a data dictionary and ask ChatGPT to group Relevant text Based on definitions found in that data dictionary.

Having a human involved throughout the entire process is key for The Marshall Project. A human-machine hybrid approach opens up new reporting possibilities without compromising editorial integrity. It helps already strapped newsrooms to overcome some of the resource constraints that they're facing while allowing reporters, designers, and product teams to prioritize their resource decisions by revealing what is a must-have now, what needs to be added later, and what shouldn't be greenlit at all.

Newsletters

Another approach that newsrooms are taking involves newsletters. ARL now has been testing the waters around some of this as reported in the Neiman journalism lab. They've been using it to digest and condense information, Create summaries and bullet points, help them simplify their research and their search processes, and even do basic tests like text correction and simple copy editing. They've also used the tools to help categorize their articles into positive, neutral, and negative buckets for potential social media locations later on. They're even using AI to create two-minute audio news summaries as part of their newsletters and to generate a thought of the day and a haiku of the day for their evening daily debrief Website posts.

Tools for Headline Creation and SEO

Moving away from the public-facing stuff, You'll also find programmers and developers that are creating tools like Yesio, which is a Slack integration by RJI fellow Ryan Rastivo. It was created to help reduce the amount of time required to come up with relevant headlines, find the right keywords to use for search engine optimization, and provide correct information at the correct time to make sure that the journalists' work is seen and Read by members of the public.

Reformatting Text

Another practical application of generative AI in newsrooms is reformatting text. Journalists spend a lot of time reorganizing and reformatting text, applying different formatting styles to the text such as markdown or CSS and HTML. Allowing these tools to work on certain sets of information rather than having them generated from scratch helps to ensure that the amount of hallucinations that are generated is reduced because it's not having to come up with any new information. It also allows journalists to save time on these tedious tasks, which would normally take little chunks of their time throughout the day, adding up to a significant amount of time wasted.

Scraping Newsroom Archives

Finally, one of the more ambitious applications that some newsrooms are talking about is the idea of using these tools to scrape through their own newsroom archives. This would provide not only a better understanding of their work internally but also allow members of the public to Interact and ask questions about their reporting. For instance, if a news organization publishes a story, typically the only options a reader has is to read the article in its entirety, check the comments section, maybe check social media, or hopefully go through the rest of the publication's website and search for other articles with similar tags. But if newsrooms were able to fine-tune some of these models on their entire story archives or even create specific embeddings from stories that have similar tags or are focused on a similar topic or issue, newsrooms would be able to allow members of the public to ask questions not just about that one story but all of the information and reporting that has been gathered and published. Using these Context embeddings to pull information from previous stories and incorporate them into the information and memory that's stored in the bot on their website would be a game-changer for newsrooms.

The Future of Generative AI in Newsrooms

Of course, these tools are so painfully new that it's almost impossible to understand or predict where we'll be even six months from now. And in fact, by the time you watch this video, most of what I said may already be out of date. But one of the most important things that journalists and publishers and anyone who's interested in the intersection of media and technology needs to understand is that these tools are much more valuable and transformative if they are understood as a new way of interacting with the information and the data on your computer. As the capabilities and accessibility of these tools Continue to expand, We Are very much likely to see direct natural language inputs translated into executable and material actions taken on behalf of the user. The ability to not only have conversations with these tools and allow them to spit back text to you but then take actions on your behalf, change code, edit and revise content, generate user and customer profiles, create entire websites or programs, and other tasks like this will be an equally if not more fundamental revolutionary change in our relationship with our machines.

Conclusion

In conclusion, generative AI is transforming the way newsrooms operate. From productivity and efficiency to editorial checks on biases and gaps, these tools are helping newsrooms to better serve their communities. While there are ongoing debates about the ethical and transparent use of these tools, it's clear that they offer many practical applications that can help newsrooms to overcome resource constraints and streamline their workflows. As the capabilities of these tools continue to expand, we are likely to see even more transformative changes in the way newsrooms operate.