The AI assistant was specifically developed for ovos play. While it is based on OpenAI's GPT models (Large Language Models), it cannot execute all commands the same way ChatGPT can.
Large Language AI models work in a way that they produce an output in human language that responds to the user's query.
Therefore, it might happen that the AI assistant responds with
"Yes, I have executed your request!"
- even though it cannot actually perform the task because the functionality does not exist (e.g., for security reasons).