I feel that such tool would add to narrowing of content diversity, because suddenly everybody would start chasing data derived trends and drop unprofitable ventures.
I think that Steemit mostly lacks a scheme for content creators to get rewards for a longer period of time, because short awarding period and low content discoverability takes away incentive to invest time in a post.
I don't think it will hurt content diversity but it will improve content quality. How can I know what my followers like if I don't have that data? I can guess which means use my brain to try to simulate the 5000+ brains of my followers but that wont work out very well in the long term.
There will be in my opinion plenty of niche topics.
OK, so this tool will simulate out that your followers or any other group is mostly engaged by Topic A. Company X also analyzes same data and discovers the profitable topic A. They have more resources to cover this topic and suddenly you become a sidebob in this topic. This repeats with every topic you find until you stick with topic Z, which is interesting to very little people thus not taken away and finally you drop your engagement in Steemit. Yes this will increase content quality, but are all side effects considered?
To me it sounds something like Google AdWords, AdSense and Search, which made impossible to beat first page rankings in many topics for simple people.
Also I believe that this is inevitable once Steemit reaches certain scale :(
Countermeasures? How about AI which would try and find quality postings in topic A (B and other high rewarding topics) and ensure their visibility to masses in such a way that both 75 level company and 25 level newcomer with a great post had a chance of same exposure and thus getting same high reward.