This is quite troubling (although I cannot say, I’m surprised). Currently, I’m not quite too sure about my stance on the broader internet-related issues, but I’ve definitely noticed a decrease in online activity on my part due to being fed up with a lot of things there. (I feel like I’m far too young to become cantankerous, but navigating the modern internet seems like a speedrun to becoming a cynic. So I’d rather not)
Lately the open internet movent (if it can be called a “movement”) is gaining a lot of traction, so I’m somewhat optimistic that things will sort itself out in the long run. In the meantime, I’m not really sure what to do.
To be honest, I disagree about what the harsh realities of the world are (to me capitalism and profit maximalisation seem the more idealistic options compared to the alternatives, which seem generally closer to human intuition), but I am quite young, and in any case that’s a different topic.
I have less a concern about the AI aspect itself but rather how it relates to some of our younger members’ safety. I’ve had to PM or talk to @tiwane to contact a few kids who were barely in high school and put way, way too much personal information about themselves. This included specific details of where they went to school, where their favorite collecting spots were. Enough information to be dangerous to them.
Knowing that it’s possible to get this kind of information from a basic google search of a kid’s name thanks to AI (without knowing they are on iNaturalist) is concerning. I think it raises some questions about what kinds of information can and can’t be shared on iNaturalist profiles by our younger members, or if there is some way to review these profiles before they are created.
While I agree that posting too much info on the internet can be dangerous, I don’t think it’s organizations’ jobs to police this (outside of what is required by law). iNat gives users fairly free rein over their profiles, and I think this is good. I wouldn’t want to give iNat my exact age so that they could adjudicate my profile content (more risk to me if iNat’s data is compromised), and I wouldn’t want iNat micromanaging my profile content.
It’s always been possible to get profile information for iNat and many other sites and similar via Google - the AI results from Google are just repackaging this info as others have noted. Bad actors generally aren’t going to need the AI summary if they are determined to do something bad - they can access the info via other means. In fact, the AI summary is probably more likely to have wrong information that could frustrate a bad actor than if they just went to the source…
Honestly, the most potentially dangerous info for most users will be their observation data itself (with coordinates and timestamps that are easily machine-readable and highly detailed). If there’s ever new onboarding, the most reasonable intervention vis-a-vis user privacy/safety is likely to have a little onboarding section alerting users to how their posts could be used to determine their personal locations and iNat’s options for restricting that data (geoprivacy) if users’ wish to use them.
Also, from a practical side, having iNat review profile content before it is approved/created/edited to reduce perceived risk for a small proportion of users seems both a) more-or-less impossible and b) not a good use of limited staff time. The only way I can imagine a small org being able to monitor millions of profiles in realtime would be to…train an AI model to monitor users’ profiles and their content. Not something I think most users would want.
I usually search my name/username to see what is online these days, but I hadn’t in a while and once I search my user up, it gave me an AI overview of my photography account (photography.emilypalah) mainly, but also mentioned that I am on iNat and eBird with emilypalah. My personal instagram is the same user as my iNat but I guess since it is a private account they didn’t include it in the overview?
It doesn’t surprise me that these overviews exist for us, but it still feels just slightly unsettling, lol. At least it seemed to just be pulling from my bios and not from any specific posts or observations.
This makes me so glad that there is no name attached to my account. If you search mkanimallover, you will find info about me as well as links, but it kind of stops there because no one knows my name. I plan to keep it that way, and thanks for this post!
Exactly. This is how search engines work, and the Google AI tool simply summarizes what the search engine has compiled for decades now.
Users of iNaturalist choose to put whatever information they care to share on a public profile/bio page. It’s not reasonable to expect any of that to stay private, and it’s not iNaturalist’s responsibility to keep it off search engines. If a user wishes to keep their private information off the internet, they need to make the decision to do so.
I agree that logging in should be necessary for viewing more parts of iNat. Maybe it’s time for a feature request…
At the very least it would probably be beneficial to make users more aware that any information posted on iNaturalist (or the iNat forum) is fully public to the internet without requiring a login, since many people have gotten used to most similar platforms (both social media and citizen science) requiring a login to view posts.
If you look at e.g. @mkanimallover’s “summary” above, you’ll notice that it goes beyond merely summarizing. Embellishments (llms are good at making everything sound like a flattering corporate résumé!), the ordering of items, human-judgment-mimicking lexicon (“expertise” “high-quality” “they are known”… based on what? by whom?) etc.
inspired by this post, i just googled my full name and looked at the AI overview to see what it would tell me about myself. it said what my job is thanks to a “meet the lab” bio my job posted on instagram, and then summarized my inat usage and other places that these observations have been used. well, it said that i contributed “over 100 observations” to inaturalist and if im being honest that offended me just a little bit. I’m at almost 180x that.
EDIT: also, from scanning and summarizing the lab instagram and uses of my observations, it gave me a lot of credit where it was not due. like, i would never use this to make a resume, and a lot of it would be just lies, but if i’m lucky enough to have a prospective employer use this to background check me…. they’ll be dazzled.
Sure. As you mention, this is how current widely-used LLMs tend to operate. That said, in these cases, it doesn’t seem to be suggesting anything fictitious or incorporating anything inaccurate information, instead just reinterpreting the text as a human might.
The bigger piece of what I’m saying here is that this is what these AI platforms do this for the entire internet. What you’re talking about is a debate on LLMs in general. This has nothing to do with iNat besides the fact that iNaturalist profiles are publicly accessible webpages.
An iNat Forum profile does not even require an iNat profile. No need to make obs, or IDs, or actually use the site. Anyone, anyone, can come here and ‘join the conversation’. (I have left iNat, but this is how I want the rest of you to do it - you must be joking. Not amused) My profile is an active identifier. Obs stacked up till I resolve my techie challenge.
Well, surprisingly, looking for my name returns information about a handome man with a great career. Although it’s not me…
Then search again another great man with another great career. Still not me.
I asked to specifically search iNaturalist, it doesn’t seem to find that really useful. But finds a young man who still has a more interesting profile.
I’m happy about the level of privacy, and yet can’t avoid feeling a bit snobbed by the search
I don’t think there’s any point leaving iNat, the information is already on the internet and all it’s summarising is what we’ve willingly put out there across different platforms. All we can do now is damage control. Obscure location on observations, take off your full legal name and use an alias, just a first name or even initials, use a profile photo that doesn’t show your full face and facial features, and then do this across whatever profiles and social media you have. I’ve been quite lucky in that I’ve kept my “digital hygiene” quite tight for the most part.
@nickupnorthuk I began blogging as Elephant’s Eye, then Diana. Then Google Plus promised us Authorship - fell for that then. Now I am more wary - but - since I have a footprint on the interwebz I prefer to add my text and influence what can be found. My obs are never uploaded in real time.
Some iNatters are working scientists - with their real name and face, some use a 'nym (because this is not my work profile) And some have very good reason for being in the 'nymwars (also Google Plus memories)
I used to search my name. Now I have to search my user name as well.
It is nice to see that my iNat photos show up as illustrations for some conservation park species lists - with credits.
I need to make sure I include an artistic photo with observations of rare plants.
That’s great! I hope you can keep it that way!
Sometimes your digital hygiene doesn’t depend on just you, though. My uni published one of my photos on their site, including my full name, and the link to the iNat observation. Since that’s now on the internet, there’s no real point in obscuring my name anymore.
In terms of anonymity, this username is “burnt”, I guess.
But I guess it was never a “throwaway” name to begin with, since most people I know irl know that “eyekosaeder” is me. And it’s a name I like, so I’m generally fine with it being linked to my real name now.
