Industry
Media & Entertainment

Core Technologies
The technology leveraged on this project included:

ChatGPT
Language: Python
Database: Postgre SQL

Project Solution

Reduced human labor requirements by 60 percent through automated processes powered by Generative AI, minimizing the need for human intervention in a continuous cycle of news content.

The client is a leading newspaper company operating in Latin America and the U.S., and it sought out Chetu's services to reach the next level in its success. The client sought to leverage AI to automate the journalistic process, including content design, creation, and distribution. They wanted to utilize pre-trained models similar to Chat GPT-3 to generate articles, headlines, and various content components.

Additionally, they wanted to generate an automated pulse for the news that would provide content for each hour of the day based on the day's headlines. The client wanted us to create an AI system with a few target objectives. We provided the staging and production server for code deployment, and the client had final approval for design choices made by Chetu.

How it Works

Chetu's developers created an OpenAI account and a script/tool that uses logic and the OpenAI API. We used a pre-trained model, ChatGPT 3.5 Turbo, that had more advanced features than the free online version.

Employees can manage system roles and prompts to ensure everything functions efficiently. The pre-trained model fetches data from wires like the New York Times and Bloomberg using its APIs. The client can generate news articles in Spanish based on the system's role and prompts.

We ensured that there was a way to manage the filtering of fetched data from the wires before sending it to the ChatGPT models to get different outputs of articles. The client can make the switch from English to Spanish for the final production of articles. They manage the initial and final versions of articles and store the input and output information.

Key Components

This project leverages different aspects of technology, primarily Generative AI. Understanding some critical components enables one to comprehend how this project functions and the scope the technology possesses. Reinforcement learning is a method of teaching AI that sets a goal without specific metrics, encouraging the model to test different scenarios rather than find a single answer. It is an especially important component because we are dealing with content that should be fact-checked to avoid misinformation.

The input source content is the raw material the AI system will analyze and reference, including existing articles, news sources, and other relevant content. Content generation modules create the articles, headlines, and content based on the input from the pre-trained model. The new pulse generation creates a "pulse of news" for each hour.

Transforming Journalism: Chetu’s Advanced AI Powers Next-Gen News Cycle

Industry

Core Technologies

Project Solution

How it Works

Key Components