Thank you for this amazing comment @crypto.piotr!

I wasnt aware that speach recognition is giving AI so much trouble. Interesting ...

Well, I rather meant that understanding the context and interpreting what was said is a problem for algorithms. Only a few months ago Google Assistant has achieved this level of advancement that it is able to for example make a phone call to a hairdresser and make an appointment. And this is considered as a huge breakthrough after a long time of development, so I suppose it is really problematic.

To create interesting publication -> it would actually require loads of creativity. Something AI lack.

I agree. Most likely such publications would be just "dry".

But at the same time it would be so easy to use AI as an "assistant". It could easily learn our writing style, analyze content on Steemit and dro short but valuable comments. That would absolutely fool majority of people out there.

I haven't thought about that, but it is really probable scenario. Many users would probably use such bots without any remorse in order to gain bigger influence and audience. And you are right, many users, especially newbies will never notice that something is wrong.

The worst thing is that such bots would be extremely hard to detect for cleaning-bot like @cheetah, as they would not produce similar content that violates the rules of spam on this platform.