Psychotherapy AI Bot

EDITH the Psychotherapy AI Bot

Inspiration

The increasing prevalence of mental health issues globally, coupled with the barriers to mental health services such as cost, stigma, and accessibility, inspired us to seek an innovative solution. With the recent advancements in artificial intelligence and natural language processing, we were driven to leverage these technologies to create a mental health support system that is accessible, affordable, and user-friendly - hence, the birth of the Psychotherapy AI Bot project.

What it does

The Psychotherapy AI Bot is a virtual companion that listens and provides empathy and support to individuals seeking mental health assistance. It's designed to understand and respond compassionately to user inputs. It provides users with the space to freely express their thoughts, feelings, and concerns. Based on the nature of the interaction, the bot can recommend relevant professional help, deliver therapeutic advice, and help users better understand their mental health. Importantly, all conversations are private and stored securely to respect user privacy.

How we built it

We started by using the Cohere platform to train the chatbot. Cohere's advanced language model allowed us to imbue the bot with empathy and understanding, which is crucial for mental health assistance. The web application was developed using flask with an addition of HTML, CSS and bootstrap for the frontend. We hosted the application on AWS due to its scalability and security benefits, which are essential for a project with such sensitive data. For secure storage of the chat history, we opted for Pinecone, which provides an efficient solution while prioritizing data privacy.

Challenges we ran into

Ensuring the chatbot responds empathetically and appropriately to a wide range of user inputs was a significant challenge. Mental health is a complex field, and ensuring our chatbot can provide quality support required careful training and extensive testing. Additionally, ensuring user data privacy and secure storage was a challenge due to the sensitive nature of mental health discussions. We worked diligently to create a safe and secure environment for our users. Furthermore, we had an issue deploying into AWS as we ran into an issue with pinecone where index exceeds the project quota of 1 pods by 1 pods and we need to upgrade our account to increase the quota.

Accomplishments that we're proud of

We're proud of developing a chatbot that genuinely provides empathy and therapeutic advice to users, which could potentially be a game-changer in the mental health space. The integration of various technologies like Cohere, OPENAI AWS, and Pinecone to deliver a seamless and secure user experience is another accomplishment we're proud of. We're also proud that we could respect and uphold user privacy, even with the requirement of personalized responses.

What we learned

Throughout this project, we've gained valuable insights into the complexities of mental health and how artificial intelligence can play a part in providing support. We learned the intricacies of training a language model to understand and respond to sensitive topics. We also learned about the importance of user data privacy and how to implement secure storage solutions.

What's next for Psychotherapy AI Bot

Looking forward, we plan to enhance the Psychotherapy AI Bot's capabilities by refining its training and increasing its understanding of mental health complexities. We also intend to incorporate a feature that enables users to schedule appointments with professional therapists directly through the chat interface. Besides this, we aim to create an app that can be active 24/7 to provide anyone with counselling they need. Lastly, we aim to collaborate with mental health organizations to bring our application to a broader audience and contribute to global mental health support.

Built With

Submitted to

Pinecone Hackathon: Jumpstart Real-World AI Applications

Created by

During the development of the project, my main contributions revolved around setting up the Flask server and taking on the role of team lead.

Setting up the Flask server involved configuring the necessary routes, endpoints, and handling the HTTP requests and responses. I worked closely with the team to define the API structure and ensure seamless communication between the frontend and backend components. By implementing various Flask features and leveraging its extensibility, I helped create a robust and efficient server that could handle the project's requirements.

As the team lead, I took on the responsibility of coordinating and managing the development efforts. This included organizing regular team meetings, delegating tasks, and ensuring everyone had a clear understanding of their roles and responsibilities. I facilitated communication and collaboration among team members, fostering a productive and positive working environment. Additionally, I provided guidance and support to team members, assisting them in overcoming challenges and promoting their professional growth.

In summary, my key contributions to the project included setting up the Flask server, establishing the necessary endpoints, and taking on the role of team lead. Through my technical expertise and leadership skills, I played a pivotal role in the successful development and execution of the project.

Mohamad Suhair Mohamad Subair
I had the privilege of leading the UI development for our ambitious project while also tackling the intricate task of integrating Pinecone, a powerful vector database. As a first-time user of such a sophisticated technology, I initially felt a mix of excitement and apprehension. However, I embraced the challenge head-on and ultimately triumphed in delivering exceptional results.

Navigating the complexities of Pinecone integration demanded meticulous attention to detail and a deep understanding of its capabilities. It pushed me to expand my skill set and explore innovative approaches.

Throughout this endeavor, I faced numerous obstacles that tested my problem-solving abilities and perseverance. From fine-tuning the UI to ensuring seamless data synchronization with Pinecone, each step presented its own set of unique challenges.

This experience has been instrumental in my professional growth. I gained invaluable insights into the potential of vector databases like Pinecone and their role in enhancing user experiences. It also deepened my understanding of UI development. Additionally, working with my team mates on this project has enhanced my learning process. To improve the UI and maximise the Pinecone integration, I actively sought input, participated in productive conversations, and drew on the knowledge of my colleagues.

Mohammed Magdi
In the hackathon, I played a crucial role in integrating Cohere, a smart language tool, into our project. With Cohere's advanced language understanding, we created a chatbot that had meaningful conversations with users. I trained the chatbot with specific knowledge, enabling it to provide helpful and relevant answers. By connecting a useful dataset, the chatbot became even smarter. Collaborating with the team, I offered support and guidance, resulting in accurate responses that satisfied users. Our project achieved its goal of delivering a smart and helpful psychotherapy chatbot named EDITH.

Ariful Islam
During the hackathon, my primary focus revolved around harnessing the potential of OpenAI, an exceptionally robust platform for natural language processing, in order to augment the chatbot's capabilities. Through skillful utilization of OpenAI's sophisticated tools and cutting-edge models, I made substantial contributions towards enhancing the chatbot's intelligence and its adeptness in comprehending user inquiries. Furthermore, I assumed the responsibility of meticulously testing the system, leaving no stone unturned in ensuring its functionality and dependability. Leveraging my expertise in OpenAI and employing a methodical approach to testing, I successfully validated the chatbot's performance and reliability, thereby guaranteeing a superlative user experience of utmost quality.

Salah Uddin