Inspiration
AI For Sensitive Security is a smart, AI-powered tool that scans email text fields in real-time to detect sensitive information such as passwords, social security numbers, and personal contact information. Once detected, AI For Sensitive Security warns the user of the information leak before it can be sent.
What it does
AI For Sensitive Security is a Chrome extension designed to enhance data privacy and protection by automatically detecting and highlighting personally identifiable information (PII) in real-time as users browse the web or compose messages. It leverages an advanced AI model from Hugging Face to analyze text content within the browser, identifying various forms of PII such as email addresses, phone numbers, names, and social security numbers. Once detected, the extension uses JavaScript to dynamically highlight the sensitive data, alerting users to potential privacy risks. This tool is particularly useful for individuals and organizations looking to maintain high standards of privacy and prevent inadvertent data exposure.
How we built it
AI For Sensitive Security was crafted using a multi-faceted approach. At its core, we utilized a robust AI model from Hugging Face to accurately identify and classify PII data. This model was integrated into a Flask application, which served as a reliable backend, hosting the AI and handling requests. The user-facing component was a Chrome extension, brought to life with JavaScript, which provided the real-time detection and highlighting functionality. This setup allowed us to create a seamless and interactive experience for the user, ensuring that sensitive information could be detected and managed effectively within the browser.
Challenges we ran into
Developing AI For Sensitive Security was a complex task that involved overcoming several technical challenges. Selecting the right AI model for PII detection required extensive testing to ensure accuracy and efficiency. We also faced difficulties with JavaScript during text highlighting, where the initial logic often led to over-selection. Back-end stability was another issue, as we worked to keep our Flask server from crashing under load. Additionally, styling the Chrome extension to look better had challenges involving the limitations set by google chrome. These challenges were significant learning opportunities that honed our problem-solving skills and resilience.
Accomplishments that we're proud of
We are immensely proud to have developed a working prototype that stands as a testament to our dedication to this hackathon. We are proud that we used a AI model from hugging face in a way to protect personal information. The creation of our first Chrome extension marks a significant milestone in our development journey, showcasing our ability to translate complex functionality into a user-friendly tool. Additionally, we take great pride in the intuitive design of our extension.
What we learned
We learned how to utilize and implement models from Hugging Face, which opened up a new realm of possibilities in AI for us. This experience has given us insights into the intricacies of machine learning models and how they can be applied to real-world applications. Additionally, we ventured into the world of browser extensions with the development of our Chrome extension. This process taught us about the unique aspects of building software that operates within a browser, interfacing with web content, and enhancing user experience without compromising website functionality.
What's next for AI For Sensitive Security
Looking ahead, we aim to diversify our approach by incorporating multiple models or fine-tuning existing models with a bespoke dataset tailored to identify a broader spectrum of sensitive information that is pertinent to our objectives. This will enhance the precision and adaptability of our tool to various contexts and data types.
Furthermore, we plan to extend the functionality of our AI beyond the scope of email composition. Our goal is to develop a versatile system capable of monitoring and securing sensitive information across all sections of an email, including received messages and stored drafts. Beyond email, we envision expanding our solution to safeguard sensitive data across different platforms and websites, providing a comprehensive security tool for users navigating the digital landscape. This expansion will not only increase the utility of our extension but also solidify its position as an essential tool for online privacy and security.
Log in or sign up for Devpost to join the conversation.