I believe that would be the thread we are currently in here. But limited to issues of how observations are obscured, and not what is obscured or why, which is a very valid but separate discussion.
I think it is important to keep some perspective here. This was an incremental change affecting only observations that were already heavily obscured geographically. In what way has the obscuration of date and time for those observations made them so much less useful than they already were(n’t)?
For those doing serious research there are still many ways to legitimately obtain the more precise data, including personal messages, using the individual trust system, and creating projects that allow people with useful data to join and share their unobscured data with project curators. Nothing has changed about these options with the addition of temporal obscuration.
I think obscuring the date of the IDs and comments specifically does more harm than good. I know at lot of people don’t immediately upload stuff right when they see it, so the upload time is pretty decoupled from the observation time which is what we’re really trying to hide.
Why is this implemented for intentionally obscured observations, though? If people are in a place that they don’t want to disclose the location, they won’t unobscure certain observations in that area. Knowing the date won’t matter if all observations they upload from that location are obscured, right?
I absolutely agree that there is too much auto-obscured stuff on the site. I understand that there needs to be consensus from the relevant community members to unobscure something, but I feel like people mix up “is sometimes poached” and “is seriously threatened by poaching” a lot which I would argue are not the same thing. A mundane example would be that people poach mule deer, but there’s no reason to obscure them all because it’s not threatening the species.
That being said, I begrudgingly think that blocking the times is probably the way to go because the alternative is to force-obscure everything observed on the day a protected species was observed and I absolutely do not want that to be implemented. I do think obscuring the comment/upload times is overboard, though, and will cause more annoyance than benefit.
Date and time data can tell a lot of information about a species even when its exact location is obscured. Very useful for things like proper occurence season, when things flower, fruit, ect. There are so many pieces of ecology that can be uncovered through observations without the need to have their exact localities.
As for obtaining yes inat has numerous ways researchers can obtain this data, except there is a considerable issue with each. Firstly mobile users (of which inat has many) will not respond to access requests/ever see them. Not every user on inat does respond when they do. Some people miss messgaes in the slew of IDs, comments, and other notifications they receive. Of course there’s also going to be individuals which simply are uncomfortable releasing data. All this combined serves to significantly inhibit researchers abilities to use data as a whole. Though this is also a problem of obscuring in general. By taking away date and time the one major piece of data that was actually usable without needing to contact every observer under the sun becomes much more inaccessible.
My main issue with all this is because it was one of the key pieces of data you didn’t have to ask for. I fully understand the need to obscure the exact locations of species for their localities. Yes not every threatened species on inat is currently impacted by the change, but it is still far too many for its utility to actually be helpful.
The important thing to note is that I still see the usefulness and utility of the feature if used correctly. By making this a add on request kind of feature we can allow those who profoundly understand the threat to specific species to make those which need it receive proper protections. If this was just that I think I would actually be elated by the feature.
Well, I do acknowledge that the number of research projects unable to use iNat data due to obscured data is probably not largely significant in the greater scheme of things. But that’s not a success in my books.
For the sake of response I also want to point out that over 40-45 personal messages, most on behalf of researchers, or just on my own, not one person has been happy to divulge locations. The folks I know who were doing regional projects who also inquired had similar experiences. And sure, no one is obligated to give it out, and this is “working as intended”. But I digress. I will consider working this side of things into a separate topic posting.
Another thing to note is that what iNat is doing here is completely in line with the ways at least one research specimen database obscures their sensitive specimen data.
In this example, don’t ask me why this particular species, or why it is obscured inconsistently by different institutions – I don’t get it either. But it shows what obscured versus unobscured data look like on that site. It’s more draconian that what iNat does, showing no date at all, and no location below County level. But legitimate researchers can still gain access to the data.
Here is a more consistent example from the same site, showing how few data are left unobscured for sensitive taxa.
And then this circles back to the old debate of “should iNat assign users as ‘legitimate researchers’ or ‘experts’, and how do we verify them as legitimate”. A solution, perhaps, but I think one that the staff are not fond of(?), either way arguably a complicated topic.
Just to give some context, internally we’re trying to separate this larger/complex issue of geoprivacy into 3 loosely coupled smaller issues -
We tried to lay out that framework in the new geoprivacy documentation https://www.inaturalist.org/pages/help#geoprivacy
This is further complicated by geoprivacy (people obscuring their own obs for all sorts of reasons and choosing how to share that information) vs. taxon geoprivacy (iNat automatically obscuring obs narrowly for ‘poaching’ concerns and iNat having more flexibility about how to share that information with the conservation community).
Re: geoprivacy, we can’t control (1) who obscures what and (3) who they share it with - its up to the observer. But re: taxon geoprivacy, we can control (1) and (3). I think all 3 are coupled in that we envision iNat as a platform where few things are automatically obscured and as a result 1 and 3 only happen rarely. However we’re still doing a lot of research and learning alot about 1 and 3. Currently, about 6% of iNat obs are obscured:
However, this thread and the changes we made are about (2) the mechanics of how iNat obscures. Many of the comments are more about (1) what gets obscured and (3) how this information is shared (I’m assuming with regards to taxon geoprivacy not geoprivacy). It would be helpful if we could focus this thread on (2) the mechanics.
Our intent is that, in a perfect world, if an observation is obscured then the mechanics should make it publicly impossible to precisely locate the observation while not impacting the access to Information about any other attributes or observations. Again, from the mechanics perspective it doesn’t matter whether this is someone’s home address (obscured through geoprivacy) or a threatened species sought after by poachers (obscured by taxon geoprivacy). Unfortunately, we’re unable to make a perfect system since (as described in the help documentation) there will always be risks of vulnerabilities to these mechanics. Likewise in the real world there are costs and other tradeoffs to consider. Some improvements are just technically complex/expensive changes to make and others start having side-effects on data we don’t want to restrict (e.g. other observations or other attributes such as dates).
We’re trying to strike that balance in the mechanics sub-issue of making it hard to re-locate an obscured obs while also dealing with those costs and open-data-philosophy constraints. We think these date changes better strike this balance but we welcome discussion about striking the mechanics balance right (2) but please lets avoid wading into (1) what gets obscured and (3) who gets access to obscured data on this thread.
I assume this 6% statistic is all of iNat’s observations, rather than only taking into consideration observations identified at species and more specific levels, i.e. those affected by conversation status.
I think really my establishing point for this topic is not whether obscuring dates helps prevent poaching, but whether it actually has enough feasible benefit to justify. As others have stated, this change adds a minor hiccup for those who are actually determined to abuse the dataset. “Death by a thousand cuts” doesn’t feel like a solution.
And my other point still stands regarding specific taxa being affected, not just any “special status” species. These changes have more impact on species that don’t need obscuring, over those that do.
This is what I am trying to get at as well. My issue is exclusively with the large number of species this change affects that gain no benefit whatsoever and reduce the utility of their contributions to citizen science.
I don’t mind these species being obscured though. I mind them being obscured via date and time, which in my eyes is relevant to this thread. Most species being obscured is something I can deal with via regular flags.
I agree with @zdanko on this one, when people intentionally obscure it’s usually private property and they obscure everything they saw. Yes when I find a nest I obscure it, but there are already a lot of ways to protect those locations without this.
One other thing, I really don’t like the fact that this also hides how many observations have been made by the observer, if I’m responding to a question or explaining an ID, knowing whether they have 20,000 observations and this was made three days ago (so I should probably include a tag in the comment to ensure they see it), or this is their only observation and it’s from four weeks ago (so no need to tag, and maybe no need to comment at all), is useful information to have.
“Why is this implemented for intentionally obscured observations, though? If people are in a place that they don’t want to disclose the location, they won’t unobscure certain observations in that area. Knowing the date won’t matter if all observations they upload from that location are obscured, right?”
Windows doesn’t wait to send you a security update until 10% of people have been hit by a bug or security flaw that allows a new attack vector. iNaturalist shouldn’t either.
It’s good that they’re being proactive. I certainly welcome the update, even if it’s a bit frustrating (yes, I’m curious where someone saw that unusual plant), but I’d hope that people here would want to put the interests of nature and the environment above their own personal preference for ease or comfort, especially when it affects the latter to such a small degree.
No, the previous geographic-only means were not sufficient. It was pretty easy before to figure out via interpolation. This, at the very least, potentially adds larger error bars on any interpolative process.
That might be the case for certain species, but I don’t think it’s a general rule. Many people do not realize that they’re posting something rare or vulnerable. They just want to know what it is or thought “hey, that’s pretty”. Outside of a handful of people, hardly anyone knows what Haller’s Apple Moss is nor that it’s endangered; to them it’s just a moss. The same is true for most other endangered species There’s about 100 species obscured for Canada and as someone with less-than-expert, more-than-average knowledge, I maybe recognize 20% and of those most I didn’t realize how vulnerable they are. Occasionally users on the more casual end of the spectrum will find something really rare and interesting. (That’s one of the things that makes the platform and IDing extra fun and exciting from time-to-time!) A platform needs to cater to that lowest common denominator of expertise, especially with regard to privacy and security.
There was another site that I once frequented that was used by a wide range of users. I, along with another user, discovered a security flaw that revealed individuals’ locations with good precision. That site failed to strip EXIF data from their photos, and users tend to post photos from within their homes often. The average user is blissfully unaware how this can reveal one’s location or home address. (This was the reason why my first concern here was what happens to EXIF data on photos which are uploaded on iNat.) What I found on that other site was that forum posters, the 1% of that community, tended to dismiss the issue saying, “Well everyone should know that! If they don’t want the info out there, they should have stripped it in advance” … But the reality is that not everyone knows their photo contains GPS coordinates. The issue affected about half of the accounts that had posted photos. I doubt that most of them realized. My takeaway was that you really need to examine what a typical user knows and does rather than the more expert ones with whom you probably associate more often (either users you follow or those on the forum).
As for what species are covered, this isn’t the place to discuss. But I’d like to point you to something else: I’ve been generally really pleased with the approach of iNaturalist Canada on this front. See: Updates to taxon geoprivacy and conservation statuses in Canada. I haven’t seen other countries discuss their taxon privacy lists on the forum, so perhaps they should model their approach after that of the Canadian team. Maybe one of the staff or moderators could set up a separate wiki thread for that topic at some point.
I absolutely want to emphasize this. People who look at the forum are not typical users, and people who post here even more-so. Personally, I really appreciate the slightly higher barrier to access sensitive data (whether for conservation or personal reasons).
A lot of disturbance of natural habitats is done by careless casual observers excited about a new thing, and it accumulates. A lot of crime is done spontaneously by casual associates. This will certainly cut down on the number of low-hanging fruit for both those groups.
