The message itself is not unexpected when there really is an unknown content type. But indeed, when it mentions a content type and magic value that involves strange characters (like in your screenshot), then this indicates that it received garbage data.
You could try out the workaround that was mentioned in the thread that you linked to, but … if the problem is originally caused by libcurl not being able to access the path with (brackets), then the same error might appear again.
Is it possible (with reasonable effort) to export the data with paths that do not involve special characters?
(Otherwise, Kevin or maybe @agallegos might have further ideas about the reason for the error, and possible workarounds or solutions…)