Oh dear...
A few years back we went to Namibia, and saw Ovambo Sparrowhawk. The guide found a pair and he confirmed the ID with playback.
My wife is still going through the photographs for the trip, and is now submitting them to an online Chinese photographic library. The other day she submitted a photograph of the sparrowhawk, which was promptly rejected by the site's AI system, saying that the bird was a Gabar Goshawk. I double checked the photos and the the system is correct! - the features are not correct for Ovambo Sparrowhawk and the bird does look good for Gabar Goshawk.
It is not that bad, as we thought we had only seen Gabar Goshawk poorly in flight on a different date, but still one to be deleted from the life lists.
The moral of the story... guides are not infallible, and don't jump to quick assumptions, particularly on a single piece of apparent evidence - thinking about it, it seems quite likely that any raptor would respond to another raptor species in it's breeding territory, so the playback may always have been a bit dubious.... and the other lesson... some of these AI ID systems are getting rather good! (so far the Chinese system we have tried seems better than Merlin!).