I’ve noticed in the last few days a huge number of falsified observations belonging to various projects set up for this year’s City Nature Challenge. There are both deliberate falsifications and lack of accurate flagging of captive/cultivated records.
For some reason I cannot seem to flag the projects involved, perhaps the functionality was limited while the CNC is happening, as with taxon changes.
Anyway, the main categories seem to be:
Massive numbers of sockpuppets,
even bigger numbers of seemingly school children accounts that verify each other’s incorrect IDs,
a large proportion of cultivated or captive organisms marked as wild,
some of these have been flagged correctly by identifiers and then overridden by observers and their friends or sockpuppets
large numbers of copyrighted images
large number of images with different dates in their EXIF data (outside the CNC range).
I hate to point the finger, but the majority of problematic observations seem to be coming from the Macau and Cochabamba CNC projects. It’s going to be a fairly large job to get these cleaned up.
I can corroborate this. I opened a flag about one particular case which is emblematic or at least representative of the Macau problems. The problem observations and sockpuppets are flooding in much faster than I am able (or want!) to keep up, even with others taking the lead or chipping in. That’s on top of the simply gargantuan numbers of good-faith misidentifications and unmarked cultivated organisms that are already washing over the taxonomic categories I check. Upwards of 90% of the powdery mildew IDs during the City Nature Challenge are erroneous unthinking selections of bad computer vision suggestions, for example.
I assume some of the projects will need to be flagged, or their admins notified at least. I’m not sure how else to deal with these nightmare scenarios en masse though. Any suggestions would be great…
Short rant that no one is obligated to read:
Honestly, this is sapping my goodwill. I’m sure plenty of good comes from the CNC year-over-year, but on top of the simply awful numbers in the incorrect “fungi” observations, the avalanche of cultivated plants is virtually taking a grater to my patience – and this was incredibly easy to foresee. How many plants in the city landscape are wild weeds versus deliberately horticulturally placed? How prominent are the tiny wild weeds compared to the large blooming horticultural plants? How many city schoolchildren understand that the street trees and flowerbed forbs are cultivated, rather than all plants being “wild”? How many children care about the iNaturalist rules even if they are aware of the distinction between cultivated and wild? I guess someone somewhere in the organisation of the event has given thought to this and concluded that it’s worth it even so. I’m sure all this has been rehashed in some thread somewhere that I wouldn’t enjoy reading. I’m just exhausted by this, that’s all…
I have an additional personal peeve. I keep on marking observations as ‘date is not correct’ because they are old photos (sometimes years old) that people are slipping in as CNC obs. Because all the projects I have encountered allow casual obs, these falsified observations will continue to be part of the CNC project.
THIS is the one that’s particularly sapping my will to continue.
Went to the Cochabamba page to see what was going on. I didn’t see any sock-puppets there but more than one pet rabbit (and other domestic/farm animal) not marked as captive. There are no wild rabbit species in bolivia, so the site marks them as invasive. Also numerous duplicate obvs. Very frustrating as some of the real obvs on there are super cool!
I’m amazed and distressed that anyone is so motivated by the CNC badge that they would outright falsify the date of their records. I haven’t seen any of these yet… but I haven’t even been looking, since the notion of their existence never occurred to me at all. Now I’ll be on the hunt for anything even more anomalous.
Yeah, most seeming sockpuppets are in the Macau one, though I’d always caution that it’s hard to separate a sockpuppet from a schoolmate who as-yet has none of their own obs but are blindly agreeing with all of their mate’s.
Ideally, we’d get confirmation from staff that they’re using the same IP, but I’d guess they’re pretty busy atm.
This reminds me of the thread awhile back discussing the gamification of iNaturalist. Unfortunately, you can only reward citizen scientists to a certain point, beyond which the consequences they will face for falsified data are not outranked by the prestige of ‘winning’ something. It is disappointing to see the cheapening of science to a petty contest by individuals who cannot see the destruction wrought by their quest for glory. While the mass-review of these observations is already an active process, maybe some sort of hold should be placed on all CNC observations, kinda like what eBird does for everything, in order to prevent any of this erroneous data from entering the whole of iNaturalist. It would be a little more work, but would stop the problem at the point of influx, as it’s more difficult to extricate these bad observations once they have entered the database.
Yeah I have to agree that this is simply to be expected when making a game or “challenge” out of inaturalist. Put any kind of reward on uploading content to any website, even non-substantial rewards like virtual badges, and you’re guaranteed to start getting flooded with low quality content, plagiarized content, AI content, etc
I’ve seen it happen with ladybugs too; a lot of people stealing photos of each other or from online while others upload AI images. In one weird case, a user captured a photo of a ladybug and another user took that photo, cut out the ladybug and photoshopped it into another photo! What!?
Are there any good ways to recognise these stolen images? My current method is to look for photos that are strangely low quality (a screenshot?), oddly cropped, or of species outside their range.
On the positive side though, since many of these new users take photos of the same individual ladybug just from different angles, I can verify that the photo actually belongs to them. Funny how that works :P
As a CNC participant, I also find the overwhelming number of low-quality observations and misidentifications exhausting. I have been actively working to correct some misidentifications within Insecta. However, Macau is a small city, and citizen science promotion is still in its early stages. Unlike other countries, Macau lacks abundant natural resources for educational purposes, and local students generally have limited awareness of biodiversity.
Despite these challenges, there are still participants who genuinely enjoy observing wildlife in natural environments and documenting their findings. Promoting citizen science is a difficult process, and I hope everyone can understand. I truly appreciate everyone’s efforts in handling various problematic observations.
I do think that in order to get a few genuinely engaged people to embrace a new tool, ultimately it often involves getting a much larger number of barely interested people to use it for a few days or weeks. The number of individuals outright causing damage, for example using sockpuppets, is quite low as a proportion of participants after all. This happens with college courses that use iNaturalist also; many problems tend to result but we can certainly appreciate anyone who becomes an avid naturalist in the process. These issues are usually bound to result when promoting citizen science, even in countries with high levels of social investment in biodiversity education.
I’m not quite convinced that the two threads have the same focus – I was under the impression that this thread was intended for curators, in their capacity as curators, coming up with ways to respond – although they seem to have converged a bit after the initial post. I suppose this was bound to happen given that the “Curators” forum category regularly gets plenty enough attention from non-curators.
On the positive side - no local project (however stuffed) can compete with the global project 27K sp and already 32% RG.
For the local projects I prefer to look at the ‘species contest’
where I would presume that the ‘true’ winner is still San Antonio in Texas - who led from the start! (4.8K with 24%RG vs 6.6K and only 3% RG! for Cochabamba)
I feel that most of the damage reported by the OP could easily be prevented if iNat enforced some measures during the CNC such as:
preventing (new) users with less than 100 ish IDs to submit an ID during the event
(preventing the pile up of wrong ID weight on obs that takes ages to fix and also some sockpuppet acc creation)
create a feature for Curators to merge duplicate/similar-ish obs (many new users dont know how to submit several pics in the same obs and they dont go back to fix it even when we patiently explain how to do so. This end up inflating the project numbers)
This feature was already proposed in several other forum threads.
These 2 alone would prevent a lot of hassle already and are quite easy to make. I never understood why iNat enforced a block on the taxonomy changes during the event. It is quite frustrating to wait until it is over to finally go back and install awesome new things people find during the event and that will also change the numbers and that we can only tell them “Oh, sorry, it is indeed a new thing but I can’t do anything about it for now”.
Other than that I feel like the CNC organizers are not very engaged when it comes to explaining how iNat works and the benefits of the CNC other than the competition. I made a journal post complaining about this aspect for Brazil event and shared it to the organizers. This year I noticed things went more smoothly, so maybe they got the message.
Anyhow, I find it quite dumb that we compete to see who has more species/obs when we are acting on different ecosystems and areas. This who-have-the-most-reving-car comparison doesn’t help science or nature at all. I’m not saying the CNC is totally useless as I’ve seen new promising users and observations arise during it, but we can definitely make it better if we reduce all that slag it also produces.
new species can still be added during the CNC, this functionality is not frozen. It’s taxon swaps and the external name services provider that are frozen
I think one of the functions of the CNC is to introduce people to their local biodiversity, so preventing new users from posting observations during the four days of the bioblitz itself would prevent new people from learning about their biodiversity. Did you mean new users could post observations after the four-day bioblitz, perhaps during the six-day ID period? Or after the entire ten days are over?
I also find CNC and the several weeks afterwards demoralizing as both a curator and iNat user. I doubt that there are (realistic) curatorial tools which would make a serious dent in this issue (unrealistic hope: Google Image Search like functionality within iNat that gives each image a score of how similar it is to another iNat image, that lets curators see close matches and flag quickly).
I agree with others who note that new user education is really the only way to prevent most of these issues at present. But the same issues seem to come up each year during CNC, despite knowing that they will occur in advance, so I despair of it changing.
I’ve spent less time on iNat during the past week than I usually do because of CNC. I feel a little bit bad about this, but I also resent feeling put upon by all the content flooding in that requires correcting which is a very thankless task for IDers/curators. I wonder how many thousands (and yes, I’m sure collectively it is thousands) of hours of time it takes from the community to address the issues caused by a relatively small number of CNC projects that don’t use iNat responsibly/engage new users responsibly?