StorySearch
Maxwell Adam: University of Melbourne. David Batonda: Monash University. Taiyeb Mustufa Radiowala: Monash University. Maia Sullivan: RMIT University. Cherline Tandra: Monash University. Mees van Vroonhoven: Monash University.
Our motivation:
With the recent surge of artificial intelligence, it has permeated every aspect of our lives. From university friends using it for assignments to parents relying on it for recipes, the power of AI and especially Large Language Models is ubiquitous! However, its potential hasn’t been fully reached, children are not properly taught its capacities, despite being the individuals that will lead us into the future. Instead of using our powerful technologies to educate our kids, we feed them short-form content on social media platforms, which has harmful effects on their attention span and mental development and may hinder their academic prospects.
What purpose does it serve?
Our primary objective in building this app was to provide young children with engaging and positive content that educates them, offering entertainment while allowing them to explore topics of interest at their discretion on their personal devices. This empowers them with a sense of control, but also has the adaptive nature of short-form content that adjusts to individual preferences. Our product does this but with an aim of educating the child, rather than feeding them mindless content to keep them using an app.
How it works:
When building this website, we openly recognised the potential of AI in shaping our world. While these technologies have made so many aspects of our day to day lives easier, we identified an area where its content-generation capabilities have not yet been seen. Educational benefits for young children. Alongside this, we considered the issues facing children in our modern world. The content that many kids consume is what is often referred to as ‘short form’, meaning it is concise, straight to the point, and ‘snackable’. This can negatively impact their attention span and is not as educational as it supposedly could be.
Our platform, StorySearch, is designed to captivate young minds through interactive and educational content. From unique stories to immersive learning experiences, children can engage with a wide range of topics curated to stimulate their curiosity and creativity. Once parents create an account for their children, kids can customise their profile and gain access to a personal and open-source library of books. Their personal library will include all the books they generate on their own… using the queries they enter the search bar. Every generated book is powered by a large language model and has unique images and text that is relevant to the query topic. In the open-source library, kids can access books made by other kids their age, who also use StorySearch. Overall, kids can learn about any topic they please in a moderated format that has a complexity which correlates to their age group.
How did we build it?
In constructing our platform, we employed a multi-faceted approach. We employed OpenAI Image and content generation to create stories books from the users' query. Through API calls to our backend, we enabled users to effortlessly save their books for future access and into a shared book library for all users. We also leveraged Firebase Firestore to associate unique book identifiers with each user account. Additionally, we implemented a Vector Database to cache similar stories and images, optimising content retrieval. On our frontend, our interface was developed using React, enhancing the overall user experience.
Obstacles that we ran into:
We ran into a couple of obstacles over the last 2 days. We spent a lot of time debugging, resolving merge conflicts, all whilst attempting to maintain our vision. The some of the challenges were ensuring robust API error handling, maintaining consistent data, including content generated by OpenAI, and optimising the Vector Database for efficient content retrieval are critical aspects to address potential bugs in the platform's stability and performance.
StorySearch also did not contain as many features we had originally hoped for, largely due to time constraints. We did, however, make great use of the time that we had. Our team laboured away relentlessly from Saturday morning up until Sunday night, staying on campus overnight to work hard to bring our ideas to life. We were also constrained by the limited weekend opening hours of the libraries at Monash. We had to improvise and find alternative working spaces and got to know each other very well along the way. However, none of this compares to our biggest challenge, staying awake.
What we will take away for the future:
Over the past few days, we've gained priceless insights and knowledge. We feel that this experience has equipped us with valuable lessons that will benefit our future entrepreneurial and industry endeavours. Through this event, we've developed a deeper understanding of the processes and challenges involved with creating products, along with effective strategies for managing them. We've enhanced our teamwork skills through this experience, learning that collaboration often requires compromises, and understanding the importance of having a clear, comprehensible vision. Nonetheless, we saw every challenge we faced as a vital lesson in our personal and professional growth.
What is next for us:
We came into the weekend not knowing what to expect as many of us had never met. When we approached the end of the weekend, we had formed a strong bond with like-minded individuals, and we are confident that this experience has forged lifelong friendships among us and that we will continue working on projects together. In terms of our app, StorySearch, we would hope to one day implement all the features we originally thought of. These features include an imagine search option for very young kids, a way for parents to moderate their child's search capabilities in a settings page, as well as some extra UI features. We also hoped to create options in the book where they can learn about specific topics and ideas mentioned in their generated book. Moving forward, we would hope to keep working on this product and try to improve StorySearch.
Log in or sign up for Devpost to join the conversation.