According to the provided text, why might generative AI tools produce inappropriate responses?
- They are trained on large datasets from the Internet
- They are not capable of understanding context
- They are programmed to be deliberately biased
- They are trained on limited datasets