The main process of training the A.I. was to plug in guidelines and ensure the chatbot responds according to those guidelines and does not go out of bounds. Let’s go over each process we went through to ensure Gemini provided accurate transcriptions.
The Chatbot proved its capabilities in terms of transcription, alt-text, metadata creation, and pattern recognition. It listened to guidelines well, although its responses were not dissimilar to how other chatbots, whether trained or untrained, would respond to the types of prompts used. For the first few trials, it repeated the disclaimer at the end of each response to a prompt. However, over time, this stopped, and most responses ended with more questions about what the chatbot could do for the user rather than ensuring its own accuracy.
The Chatbot proved its capabilities in terms of transcription, alt-text, metadata creation, and pattern recognition. It listened to guidelines well, although its responses were not dissimilar to how other chatbots, whether trained or untrained, would respond to the types of prompts used.