What people ask us most often, answered without beating about the bush.
It will change many jobs, and some tasks will disappear: it would not be honest to deny it. What we see every day is that AI does repetitive, well-described tasks well, and does badly at anything that needs knowing the people, the place and the trade.
Someone who knows what is needed and can judge the result becomes more useful, not less.
Not blindly. An AI can say something wrong in the same tone it uses for something right. Treat it as a fast, well-read colleague who is sometimes mistaken: for anything important, you check.
That is why, with us, a person reads and approves everything before it reaches customers.
It grasps the meaning of sentences very well, even when they are written or spoken loosely. It has no experience of the world as a person does: it has never stood in a shop or been in a hurry.
The more context you give it, the better it answers.
That depends on the service you use, and it is always worth asking. In the conversation on this site, the sentences go to the model that writes the replies and we do not keep them.
The details are on the Privacy page.
It is a powerful tool, and like any powerful tool it can be misused or used carelessly. Today's real risks are practical rather than science fiction: mistakes taken as fact, decisions handed over unchecked, data given to those who should not have it.
One simple rule reduces them: a person decides, and answers for it.
With something small and tedious in your work: a text to tidy, a table to sort, a question you cannot find the answer to. Try it, look at what comes out, correct it.
Or talk to it for two minutes, here on the site.
The quickest way to form a view is to try it.