Polly the Parrot

Talking AI companion/pet to keep younger demographics engaged and away from screens when parents cannot spare time in this busy world.

Comment

Polly the Parrot Companion

Polly the Parrot is a fun plush that you can chat with and ask many questions! Tailored to speak with younger audiences and keep them company in an educational and engaging way. Polly keeps your young ones learning new things and staying social.

Problem

Now more than ever, we are busier in our lives. Especially as the younger generations grow up and start families of their own, it gets increasingly difficult to spare time for everything and everyone. Unfortunately, we have seen a rise in young children who cannot get away from their devices. To counteract this, we created Polly the Parrot. This companion provides the engagement and responsiveness that developing minds should have. It’s more than just a chat bot or a pet, it’s a friend.

Inspiration

Since this year’s theme was all about pirates, we wanted to go for something on theme. As we were talking about this companion, we thought about how pirates are depicted as having parrots on their shoulders. Pirates always have a close bond with their parrots and often have them as their assistants. From that idea, we created Polly, as a child's learning assistant and best friend.

How we built it

By utilizing the power of both OpenAI’s ChatGPT and IBM’s WatsonX, we were able to record our speech and have the AI’s respond in a proper manner. OpenAI's ChatGPT creates fluid and natural conversation while retaining memory of past responses to improve in the future. WatsonX was essential for the clear speech to text translation of the child's query and was similarly used to produce an audio output from the ChatGPT answer. Furthermore, we trained our response AI to have the mannerisms of a pirate to align more with the theme of a pirate’s parrot. We also implemented many open source frameworks for Python to ensure everything can run as smooth as possible. For the parrot itself, we bought a stuffed plush and put 2 mini servos in the beak that run in harmony to simulate the parrot speaking. We connected the servos to an Arduino Uno which we ran test cases in C and then converted the main code into a Python file for continuity.

Challenges we ran into

Our biggest challenge by far was getting the AI integrated with our code. Neither of us have worked with AI or APIs before this hackathon, so there was a great learning curve for getting everything to integrate. The next biggest challenge was combining the web based abilities of the OpenAI and WatsonX software with the independent Arduino framework and patching issues that came with cross-platform code in two different languages.

What we learned

This project taught us many useful skills. The biggest being how to integrate AI into coding projects. What surprised us the most however is how many resources Python has for anything. If we had any issue or feature that we wanted to include, there was always something out there to solve our specific problem. The next major lesson we learned was cross platform connections and usage of standard communication protocols like Firmata to write custom firmware to manage various devices.

Future Ambitions

Eventually, we want this bird to be a fully realized design. All the electronics run locally on the Parrot's processor, the ability to wear the parrot on your shoulder and an app that lets you take care of Polly as if it were a real parrot.

Share this project:

Updates