We appreciate the thoughtfulness of this concern. Our response to everyone is -- please be thoughtful with each classification you submit, but definitely there's no need to stress. Each image, audio clip, video file, or data file is classified by multiple people. The algorithms we run to create the final consensus result take into account all these classifications. For example, in SnapshotSerengeti.org 25 people classify each image. In a study comparing the consensus result with experts, the consensus result agreed with the experts 97% of the time. That's a huge rate of agreement. And for the 3% of time when it didn't agree, half of those images the experts don't agree amongst themselves.
The most important thing is to classify, not to be 100% right 100% of the time.
Also, if there's a subject you're genuinely unsure about, you can always discuss it on the project's 'Talk' discussion forum.
Thank you for your contributions and participation!