Not certain if this has already been posted about here, but iNaturalist recently uploaded a blog post stating that they had received a grant from Google to incorporate new forms of generative AI into their 'computer vision' model.
I'm sure I don't need to tell most of you why this is a horrible idea, that does away with much of the trust gained by the thus far great service that is iNaturalist. But, to elaborate on my point, to collaborate with Google on tools such as these is a slap in the face to much of the userbase, including a multitude of biological experts and conservationists across the globe.
They claim that they will work hard to make sure that the identification information provided by the AI tools is of the highest quality, which I do not entirely doubt from this team. I would hope that there is a thorough vetting process in place for this information (Though, if you need people to vet the information, what's the point of the generative AI over a simple wiki of identification criteria). Nonetheless, if you've seen Google's (or any other tech company's) work in this field in the past, which you likely have, you will know that these tools are not ready to explain the nuances of species identification, as they continue to provide heavy amounts of complete misinformation on a daily basis. Users may be able to provide feedback, but should a casual user look to the AI for an explanation, many would not realize if what they are being told is wrong.
Furthermore, while the data is not entirely my concern, as the service has been using our data for years to train its 'computer vision' model into what it is today, and they claim to have ways to credit people in place, it does make it quite concerning that Google is involved in this deal. I can't say for certain that they will do anything more with the data given, but Google has proven time and again to be highly untrustworthy as a company.
Though, that is something I'm less concerned by than I am by the fact that a non-profit so dedicated to the biodiversity of the earth and the naturalists on it would even dare lock in a deal of this nature. Not only making a deal to create yet another shoehorned misinformation machine, that which has been proven to use more unclean energy and water (among other things) than it's worth for each unsatisfactory and untrustworthy search answer, but doing so with one of the greediest companies on the face of the earth, a beacon of smog shining in colors antithetical to the iNaturalist mission statement. It's a disgrace.
In conclusion, I want to believe in the good of iNaturalist. The point stands, though, that to do this is a step in the worst possible direction. Especially when they, for all intents and purposes, already had a system that works! With their 'computer vision' model providing basic suggestions (if not always accurate in and of itself), and user suggested IDs providing further details and corrections where needed.
If you're an iNaturalist user who stands in opposition to this decision, leave a comment on this blog post, and maybe we can get this overturned.
[Note: Yes, I am aware there is good AI used in science, this is generative AI, which is a different thing entirely. Also, if you come onto this post with strawmen or irrelevant edge-cases I will wring your neck.]
tangentwithtanner
prguitarman
izzy806
dededos
origamiaround
snowghoul
