Generative AI to help authors write news articles

Forget frantic data scrambles and quote hunts! I’m excited to see Generative LLM models reshape news reporting. Imagine: AI spinning raw weather data into captivating narratives in seconds, complete with expert quotes and impactful storytelling. Based on historical baseball data, writing a current baseball event or season or a topic related to politics.

I built this prototype on Amazon Q showcases this magic, turning weather data into engaging news articles. No more tedious tasks — AI does the legwork. While AI won’t replace journalists, it empowers them. AI handles the heavy lifting.

The prototype feeds on:

• Real-time weather data: temperature, precipitation, wind
• Historical weather data: past trends, anomalies
• Sample weather data: for testing and refinement

Here’s what AI can do:

• Craft captivating headlines
• Paint a vivid picture: Describe the impact on the city, potential dangers
• Offer expert insights: Include quotes from meteorologists
• Highlight disruptions: School closures, transportation chaos, power outages
• Share resident experiences: Personal anecdotes add emotional impact
• Provide safety information: Official warnings and precautions
• Forecast future conditions: A glimpse into what’s next

But can AI replace human journalists? Not entirely. Human expertise in crafting narratives and interpreting data remains crucial. However, AI can free up journalists’ time for deeper analysis and storytelling. This prototype showcases the potential of AI in news reporting. As technology evolves, expect even more sophisticated and nuanced narratives emerging from the world of generative AI.

Several leading tech companies such as Amazon, Google, OpenAI etc are developing tools to streamline content creation, leveraging the power of Gen AI. This prototype uses Amazon Q, a generative AI assistant designed for enterprise use. Imagine uploading your historical data (like weather trends in this case) and receiving compelling, data-driven narratives as output.

This prototype uses real-time and sample weather data to demonstrate its capabilities. Think -20°F with 15% precipitation, 70% humidity, and 8mph winds. Coupled with historical data (date, precipitation, temperature highs and lows, wind, and weather conditions), it empowers the AI to generate a news article tailored to the specific scenario.

By configuring historical data as “retrievers”, I could prompt Amazon Q to generate a News article.

“Historical weather is in the uploaded file. Write a news article with 1000 words for a weather report. Current weather in seattle is -20F and precipitation is 15% and it is extremely cold and snowing. Write a captivating headline, Provide a brief overview of the current weather situation and its impact on the city. Mention the specific extreme condition (e.g., heavy snowfall, heatwave, hurricane) and its potential consequences. Offer detailed information about the current weather conditions, including temperature, wind speed, and any unusual phenomena. Write about historical weather when it was extreme cold and write next 7 days outlook forecast of weather.”

Amazon Q then generates:

“Arctic Blast Grips Seattle in Record Cold”

Seattle is in the grip of its coldest weather in years as temperatures have dropped to -20 degrees Fahrenheit. Heavy snow has been falling throughout the day, with over 5 inches already on the ground. The snow and bitter cold are causing major disruptions around the city. All local schools have closed, and many bus routes have been canceled or delayed.

The National Weather Service reports that this is the coldest temperature ever recorded in Seattle for this date. The previous record was -15 degrees set back in 1950. With the wind chill, it feels like -30 degrees outside. The heavy snow has made roads treacherous and caused dozens of accidents around the city. City crews have been working around the clock to plow and treat the roads, but sidewalks and side streets remain difficult to navigate.

Looking back through historical weather records, some of the coldest temperatures on record for this area occurred in January 2013, when the mercury dropped to -7.1 degrees Fahrenheit. January 2015 and November 2014 also brought periods of frigid cold, with temperatures below zero. Meteorologists say the current cold snap is forecast to continue through the weekend before slowly moderating early next week. However, lows will still drop below zero each night. More light snow is possible each day as well. Residents are urged to only travel if absolutely necessary and to check on elderly or vulnerable neighbors. Relief from the bitter cold is still at least a week away.

The bold text illustrate how generative AI effectively incorporated the data I provided. This demonstrates its potential to craft even richer and more engaging content when it has access to additional data sources. Imagine feeding it real-time quotes, current emergency information, and diverse perspectives related to the weather event. Such an enriched dataset would enable the AI to weave a more nuanced and impactful narrative.

This prototype serves as a glimpse into the future of journalism, where generative AI empowers authors to create highly engaging content. It doesn’t aim to replace journalists’ expertise but rather augments their capabilities, freeing them to focus on deeper analysis and human-centric storytelling.

Data Sample:

One potential application for News organizations for this could be by building efficient workflows when an author begins by inquiring about the current weather in Seattle, incorporates that data into their notes, and continues gathering details on historical weather, expert quotes, and any other information required for the article.

The possibilities are endless with products like Amazon Q with RAG built-in — Retrieval Augmented Generation (RAG) is a natural language processing (NLP) technique. Using RAG, generative artificial intelligence (generative AI) is conditioned on specific documents that are retrieved from a dataset or Google’s Gemini, allowing you to upload enterprise data and create sophisticated applications.