Final Project - Future Journalist

We were thinking of ideas for our project, for what the core behind our idea should be and what lessons we found to be the best from the course. In it was a desire to implement both the artistry and technicality of what we learned and liked. Initially, our idea was very broad. We had a broad thought of wondering how the news industry could be changed by AI in the future, since it is faster and more convenient than a human as well as can be constantly reading information from online databases such as social media and websites, and write an editorial-quality newspaper by trading the algorithm with tons of well-known newspaper such as the NYT, WSJ, etc.

Actually, this is still more of a polished version of our thoughts. Our initial idea for the actual project was a bit ambitious though. Our initial project concept’s aim was to combat fake news, merging interactive storytelling with user-generated content. Using multi-angle cameras, we capture audience photos using the webcam, transforming them through 'img2img' on Stable Diffusion to create dynamic, beyond-captured actions. ChatGPT-generated negative fake news are interwoven with user images for a compelling storytelling experience by placing audiences into realistic crime scene pictures.

We emphasize impact through visually striking effects and alert systems that highlight how fake news generated by such as Deepfake could bring harm to individuals like cyber bullying, etc. The objective is to actively engage users in understanding the creation and consequences of misinformation, fostering critical thinking in the AI age.

We actually thought of this idea of AI Operated News from the idea of “fake news” that was supposed to show how AI image generation could generate negative pictures based on the user's face and generate fake news based on generative AI. Then after discussing this extensively and how to best implement it, we think even though the previous idea is more interactive, we came up with a new idea that showcases more illusion and future prospects for the impact AI could have on the news industry.

With this thought, we still knew AI in the news industry was the obvious choice and an interesting one considering it would take over tedious tasks and be able to combine and compile data that a journalist would gather offline to be able to feed into the machine.

Our main point was to make journalism more convenient and enhance the process for both journalists, papers as well as consumers by making compiling and generating easier for the writers, distribution easier for the paper and efficiency at receiving news better. There is, after all, a reason as to why exactly fast media such as social medias like twitter is taking over traditional journalism, due to its speed and fast availability. With AI the news industry would be able to achieve this advantage as well.

Operations was another thought and in terms of the supply chain, ideally the AI algorithm could also directly deliver the news digest to subscribers 24/7 online or deliver a printed version that is printed in an automated pressroom. The main thought behind this initial idea was regarding the context that there is so much information going on everyday, so a news digest tool that could be editorial quality, updating 24/7 and as comprehensive as possible could only be done with the assistance of AI.

https://drive.google.com/file/d/1mqka8Q8Yl5VqbVmSM8CMb_iISw_desHp/view?usp=sharing

The project is mainly written in HTML, JS, and CSS. The basic way it works is that the program will repeatedly send requests to OpenAI API, get the content of each news, and then send it to SD to generate pictures.

As for the CSS file development, there are few technical issues along the way such as how to automatically refresh the webpage with the newly generated news constantly. And moreover, we couldn’t figure out how to adjust the format of the website for specifically the print version into an A4-friendly format nor how to make it automatically print.

When working on displaying the project and the newsroom, we received feedback that our concept and the idea itself had a big disconnect. Initially, we had a concept video that we displayed with the project.

This concept video did not really give the essence or the vibe of what our project was really about. So, we took the advice and shifted our focus to the project itself. Instead of letting a video speak for our project, we let the project speak for itself. Thus, we displayed our project on these gray blocks which kept up our props and had initially made the display a bit messy, with crumpled up paper, notebooks and a mug, to make it a ‘raw’ version of a newsroom, and more authentic. But upon consideration of the feedback, we realized that it would look better aesthetically and conceptually to neaten it up, and thus we just left the mug on the blocks as well as our printer and laptop, to give a more polished feel.

The importance of AI News Agency is that it has greatly saved time, energy compared to its traditional counterparts, it keeps sorting the numerous information into a user-friendly digest. Our project is basically a prototypical website of our imagination, using Stable Diffusion, OpenAI API and HTML CSS coding. The website will keep refreshing for new news. I think this project was something interesting and could be very useful for the industry.