AI can return bad data. A lot of people find it harder to believe. But the data on internet can be false and also AI is struggling a lot working with the right data. As it crawls the web often, the less it has accuracy on data. It has no models to verify the data. But we can make it establish some of such models.
You have to learn to ask the right set of the questions. This way AI gets trained and slowly you reach where you want to. And that is what some of the times we want. As it can bring better results for us.
I once asked the AI to write on something then I wanted it to check for the grammar mistakes. And then I asked it to check the data for accuracy. And slowly one after another you are going to be closer to the results. And that is what we want. Something you have to consider there as well.
Make the AI to do the changes. If it can do that good and if it can't do that find out why. Most of the data that it does return is kind of fun and also it can lead to some really good changes when it is corrected as the time goes ahead.
It does not matter if you are using the chatGPT or the huggingface AI. All you have to is find out if the data that you offer is correct or not. And then based on that make the changes. If you can do that better way it can be good for the output that you produce as well.
AI is powerful and we have no option than to accept it. Once we accept it then we can build on the top of the same. Slow and steady we can use it to make things. And that should be default use of the AI where things get usable for most of us.
In short, you have to add the data, you have to ask it to change the data. When you do small changes add up and then those changes eventually make sense too. Each of those changes eventually get them to work in the due time. Ai is going to be our new butler.
