Lately I've been thinking a lot about how I want my children to interact with AI, and I realize that there's not much tooling around parental control or behavioral control of the AI. I was thinking, for example, to have my own GPT with some prompt engineering to make sure the LLM doesn't give answers directly and instead acts more as a guide, and also to make it understand who it is talking to. Has any of you done something about it? I assume AI adoption in households is increasing exponentially.
I don’t use AI directly with my kids, but I’ve introduced it to my younger nephews and niece (all under 10 years old). I’ve shown them how to use ChatGPT similarly to how I used to teach them to search for information on Google.<p>Because the responses can be quite long or complex, I guide them on how to prompt ChatGPT for simpler, shorter answers. I see it more as a tool that I’m teaching them to use, like I would if I were introducing a calculator. I also emphasize using the word “research” in their prompts, especially when it’s for factual topics related to school.<p>Interestingly, while the older generation in our household doesn’t use ChatGPT much, the kids found out about it from their schoolmates. They’re familiar with it but not yet using it to its full potential. I encourage them to experiment safely, and I also suggest using Perplexity as an alternative since it’s more grounded for research purposes. Plus, they’re using my ChatGPT Plus subscription, so I try to help them maximize the experience without getting overwhelmed by too much information.