The blog post and other comments clearly state that the intention is to use generative AI to provide summaries of why a potential ID is suggested by the iNat CV system, at least to begin with. It’s absurdly disingenuous to ask “if what is implemented” when that’s been one of the stated objectives from the initial posts on this issue. It’s such a clear thing that this response has be a bit baffled and even more skeptical than I initially was.
The issue of “why so little faith” has already been discussed over and over again here, along with papers to support the skepticism. But to reiterate, some of the reasons why for “so little faith” have to to with the well documented unreliability of LLMs as a tool, the long standing stance of iNat as a ‘people first’ platform, and the instance that iNat does not use that particular type of AI. And let’s not forget all the discssions, of iNaturalist’s carbon footprint, and energy use where generative AI came up as a topic and iNat assured everyone that iNat was not using generative AI as we all know the absurd amount of energy generative AI uses.
Quite literally on almost every level this proposal goes completely against both the stated ethics of iNaturalist and what iNaturalist has said in the past.
We have all been wading through the rampant enshittification of damned near everything online and every online platform, and have been watching the most popular search engines collapse into nothing more than bullshit aggregators as exactly this sort of AI implementation has been rammed down ever user’s throat without any option for feedback or say.
Can you really, honestly, say that anyone is surprised that there is vehement backlash when the same is proposed for a platform that has consistently stated that human interaction and connection is a critical aspect from the start? Is there really anything startlin about a lack of trust in this when the proposed roll-out was so badly botched it was reminiscent of Apple’s ‘our AI gets to use everything you produce and none of you get a say’ actions (and before anyone whinges about iNat not using the data in that manner, that’s missing the point, the point is about the user interactions)?
This is both a PR disaster and an implementation disaster for iNat, and if iNat ignores its user base it’s going to devolve into just one more purveyor of LLM bullshit, instead of one of the most respected and used online platforms for scientific and nature engagement on the entire internet.