From a statistical point of view, there's very little difference between the self-described "Professional Pilots" - 81.4%/18.6% - and the "non pilots" - 76.7%/23.3% - which would tend to suggest that even if a significant number of the self-described professionals are not, they would not greatly skew the results.
For example, suppose that of the 1931 "professionals", half are really not. Based on the "honest" non-professionals, that 1931/2=965 people should have split 0.767/0.233, or 740 "unsafe" and 225 "safe". Eliminating those from the sample on that basis gives:
"real professionals": 831 "unsafe" and 135 "safe" - a 86%/14% split.
In fact, as you can see, the more "fake professionals" you assume, the worse the resulting opinions of professionals ends up being....