Like many technologies, I find AI (at it's present level) to be quite useful for somethings but I'm dubious about others. For example, I sometimes need to find things in lengthy legal documents or contracts without knowing specific words or terms to search for. An AI analysis will pull out phrases or sections that are relevant to my questions so I can then apply my human intelligence to the situation I am researching. A super-intelligent search engine, in other words. It's also great for comparing two documents or images to see what changed and accurately describing those differences in a finished form. But when I use the same tools to analyze a much broader area, e.g. the entire internet, I have to worry that it has digested errors or unsubstantiated opinions along with solid fact and is glibly quoting nonsense to me. As Ray said, "garbage in, garbage out". That's not a new worry, but AI skills make it faster and more plausible, leading to sometimes gross errors.
AI artistic creations are often amazing in their quality, making them seem credible even if utter fantasy. A tool ripe to be misused by those with an ulterior motives or a simple childish desire to see how many people they can fool. Sadly, there are an awful lot of gullible people ready to play the fool in that game. So far we hear/see more misuse than benefit on this score cause that sort of sensation is what the media thrives on.
I can see a lot of AI value in writing computer software, supplanting a set of tools/skills that were previously known as report generators and database query languages. This is the grunt work of software in business, where there has always been a huge backlog of desirable projects simply because there weren't enough skilled coders to build & test it all. A more intelligent tool can cut hours and days from those projects.
