When to assign species-level ID?

Hello!

I recently discovered iNaturalist and I’m in love. I’m also confused. I am a casual observer of nature with a good camera. Aside from some aquarium fish, I lack deep knowledge of most taxa*. Right now, when I post something I don’t know (e.g. a wildflower) I look through the AI suggestions, and if one of them matches I pick it. If none match or multiple ones match, I pick a broader taxon.

Is that right, or should I avoid picking species-level IDs unless I’m a specialist in this organism? The guides I’ve seen seem to differ. This tutorial says to be bold https://forum.inaturalist.org/t/identification-etiquette-on-inaturalist-wiki/1503 which seems to validate my approach. A starting ID will generate discussion and eventually bring clarity.

However I also see threads concerned about AI-powered misidentification by non-experts as a large scale issue https://forum.inaturalist.org/t/dont-use-computer-vision/13327/3

So what should I (and other newcomers) do? Pick the best match from the AI suggestions if one appears unambiguously better? Or give general taxa unless we have expert knowledge?

Thank you!

*and even with the fish I’ve been surprised. I was about to go proudly identifying Corydoras Paleatus. Then I saw Corydoras longipinnis for the first time

4 Likes

Hi Welcome to the forum. I prefer to add a broad because the CV suggestion can be wrong many times.

3 Likes

welcome to the forum @spaceman98 :)

If you are confident that the AI suggestion is correct, there is nothing wrong with selecting it. If it ends up being wrong, that’s completely fine; it will be corrected, and you’ll learn something new.

If you’re unsure about things, then just do exactly as you noted; ID to the taxonomic level you’re most comfortable with.

There is no right or wrong way to go about things; whether to pick certain suggestions or which taxonomic level to ID to varies depending on the context.

What is generally criticised in other forum threads re AI suggestions is when users ‘blindly’ pick AI suggestions, i.e. without checking things like distribution or seasonality to see whether those suggestions are feasible (in addition to morphology of course).

11 Likes

You don’t need to be a specialist, but you should be confident that you’re correctly recognizing the species in question.

Often if I think I know the species, but am not confident enough to give a species level ID I’ll ID to genus and include a comment that I think it’s XY species.

11 Likes

I think the approach you described is generally good but would add two things:

  1. When you ID boldly, it’s important to stay active with your observations. If someone else is able to ID differently than the AI or your initial ID, don’t worry! Withdrawing your initial ID is totally fine, or changing it to a higher taxonomic level, etc. There’s no shame there!

  2. But in the same vein, don’t blindly change your ID to agree with someone else’s ID, even if you think that they are an expert or may know better than you. This is another common issue, with users feeling like they should agree to thank another user for their ID, etc. Your ID should represent your own best knowledge/experience (not your assessment of whether another user is an expert or not). If another user’s ID or comment lets you grow your own knowledge or independently verify an ID (with a new character or mark you hadn’t known about before, for instance), then updating your ID to agree is great - just don’t do that reflexively.

A lot of the issues with the AI come about when the OP uses it, and then another user confirms without any independent expertise and the observation becomes RG. One other approach you can take is to note in the description if you’ve relied solely on the AI for ID. I sometimes do this for my observations of taxa well outside my expertise to let other IDers know not to weight my initial ID to highly!

15 Likes

I agree with everything said so far. I would also add a couple of things. Do not be afraid to ask questions. I try to leave an explanation, especially with id changes, but sometimes not.
Initially, pick a region and a taxon and stay there for a while. When I started, I tried to identify moths from all over the world, and it was frustrating! It wasn’t until I settled on Canadian Noctuidae that I started to learn. Use the identify tab to filter observations.
Leaderboards can help you identify people who may be able to help out, but they are not always reliable. I’m at the top of some leaderboards and sometimes I don’t remember the species features.
Find some resources. As you have already found out, there can be a lot of variation within a genus, and within a species.
Or not - if you are comfortable observing nature and not going ‘deeper’, then that’s cool too. There is no right or wrong way to use iNat, but it’s a fun place and a good place to learn to whatever level you want. Personally, I like the identification part better than the observing part, but that’s been my choice to make!

4 Likes

Choosing to agree with a finer identification than you have made is a bit of a judgement call. If you personally know the person and rely or trust them, then maybe it is easier or better to do than a seemingly random person.

Likewise a quick check of their profile can give some indication if they have written anything. For example the other day, I entered and identified a bee to a genus level. I was fairly certain I had the genus right. Another person came along and added a species level ID (in the genus I had identified). I did a quick check of their profile, turns out they are a PHd candidate studying the genus. Say what you will, but I was comfortable agreeing with their assessment.

5 Likes

I think the “be bold” statement might need a little fine-tuning.

Thank you all for the excellent and insightful answers!