I checked from my observations, the notes are synced. Don’t worry too much, even if some are goofy, anyone reading them, even researchers using GBIF, would be human anyway.
But if you don’t want some notes to not leave iNat, add it as comment instead of notes. All comments are not synced out for better or worse.
Also apart from notes, the other thing that is synced is first species identification’s comment subtext on that observation.
Interesting, thanks! And I appreciate your input on the notes - that’s helpful. As long my intentions are good, and the notes are intended to be at least a little helpful, I’ll feel OK about using them. 
From the beginning of this conversation, I’ve been thinking of an observation of mating beetles (or was it cicadas?) I saw quite a while ago, shared by a new user, with the note “they wuz f—in.” They weren’t wrong…and maybe it’ll someday provide a moment of levity for some researcher browsing GBIF records.
I don’t think personal comments or notes are any problem. I think the issue is just that the content should be truthful and related to what the observer saw/experienced. If someone wants to put jokes in, that’s fine (as long as they don’t violate Community Guidelines).
I occasionally come across observations from school projects where students seem to have pasted the whole Wikipedia article (or some similar source) on the species in question (or what they think it is) into the Notes. In cases where there are multiple examples in a row from different users, I imagine the instructor has advised them to do so for one reason or another, but I’m not sure why they’d do that. I guess theoretically it means the student has read something informative about the species…
For these instances, I still advocate for a third option under the annotation for “sex” - male, female, and “yes”. 
Either that, or perhaps the instructor suggested they “show their work”, as in, note how they arrived at their ID. Then the student decided it might be a good idea to copy the text into the observation, so that the text supporting the ID is right there in the observation. And then, monkey see/monkey do leads to their classmates doing the same.
I hope the instructor isn’t advocating this protocol, but who knows.
Same here. I probably have a ton of useless notes since I started on iNat by importing observations from Flickr photos and the descriptions ported over from Flickr were not written with iNat in mind as I didn’t even know iNat existed.
To illustrate the sort of notes I dislike without “outing” anyone, I prompted Gemini “summarize in 30 words how i identified this photo as orthosia hibisci” and it responded:
“The identification relied on the grayish-brown forewings with distinct spots (orbicular and reniform) and lines, often accented by vague blackish markings, characteristic of the Speckled Green Fruitworm Moth.”
I see these sorts of descriptions in notes a lot when I’m identifying. This is truly slop. “Grayish-brown forewings with distinct spots and lines, often accented by vague blackish markings” describes half the moths in the world and adds no benefit to the observation at all. It certainly doesn’t support the species-level ID given. And I never even showed it the picture; it might have been a photo of something that looked nothing like the AI’s description. If anything, the only value in such a note is that tells me this observer has no idea what they’re talking about.
Where I can see this becoming a major problem is if the AI text adds fake information about the circumstances of finding the organism. One of Gemini’s responses to a prompt I gave about how I identified Manduca sexta read:
“I knew it was a Carolina Sphinx because the caterpillar was found feeding on a plant in the Nightshade family (Solanaceae), such as tomato, tobacco, or potato”
I never told Gemini where I found the caterpillar. It just made that up. I might have found it on any plant, and now I’m misreporting the host based on my misidentified species name.
This isn’t to say people don’t make honest mistakes about host plants, etc. And I’m generally not against the productive use of AI. But pasting a block of text from a LLM and passing it off as your own ID thought process and/or context for an observation is not an honest mistake; it’s literally lying.
These seems like it’s in a grey area where because it’s not directly related to media evidence or metadata, and judgment is subjective, it’s not clearly deserving of a DQA knockdown. But on the other hand it could realistically cause issues downstream if host information or other details are inferred from the observations notes by a naive researcher. I can’t think of a good solution other than an observation field with the meaning of “suspicious observation notes”.
iNaturalist does have the guideline " Add accurate content and take community feedback into account. Any account that adds content we believe decreases the accuracy of iNaturalist data may be suspended, particularly if that account behaves like a machine, e.g. adds a lot of content very quickly and does not respond to comments and messages."
I don’t see why mass fabrication of details in the notes wouldn’t fall under that? I think there is difficulty in proving notes are fabricated without the notes being obviously wrong, and it’s likely just not really enforced in relation to notes.
For now people should actively use this observation field - https://www.inaturalist.org/observation_fields/148
It helped me in IDing new undescribed sexes confidently with projects using that field such as https://www.inaturalist.org/projects/mating-moths (ofc there are some cases of interspecies sex in lep lol, so more observations needed for confidence)
Good to know about that, thanks!
The host information is often crucial for identification. For example, for plant mildews. Fabricating the host info will fabricate the identification of the fungus.
Let’s not be petty minded, let’s simply accept that half the moths in the world almost look the same and if there are differences, who cares? The world is in permanent change, so why care for it today if tomorrow, it may be gone and its no longer an issue?
Consume and enjoy!
From a policy perspective or from a practical perspective of enforcing rigour?
To the latter, the answer is easy: There is no process or support to systematically do it and any attempt asking for it is debated away in this very forum.
I’ve long thought that the annotation Sex > Yes would be an elegant solution! Ostensibly, we’re supposed to duplicate the observation and annotate for the male and the female, and then (ideally) add links between the two…but imagine how handy it would be to just click that Yes.
I use https://www.inaturalist.org/observation_fields/6637 to annotate mating pairs, but it is “bug”-specific.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.