Does kleptotaxonomy exist? Could AI do it?
Robert Mesibov
I sent a draft to the late Lucien Koch, a Western Australian arachnologist. Koch was horrified that I was going to publish data that would allow someone else, called here a "kleptotaxonomist", to formally publish the new species without ever seeing a specimen or designating a type.
I wasn't sure at the time if Koch's fear was justified because I wasn't familiar with the Code in 1986. As it happened, two of my species were later described by chilopodologist Greg Edgecombe, who referenced my book but based his descriptions on real type specimens.
I'm now wondering about online images of new species. Section 72.5.6. says "In the case of a nominal species-group taxon based on an illustration or description, or on a bibliographic reference to an illustration or description, the name-bearing type is the specimen or specimens illustrated or described (and not the illustration or description itself)."
There are millions of images of animals online, some of them new species. Could a rogue taxonomist publish new names and descriptions for existing species based on high-quality images of minor variants? Could a "kleptotaxonomist" publish new names and descriptions for new species based solely on an image?
More worryingly, could a "kleptotaxonomist" get an AI agent to find possible new species images online, generate plausible new names and descriptions and get them formally published by the "kleptotaxonomist" assisted by AI?
This is all possibly way too unlikely, but I'm curious to know if there are real, past examples of "kleptotaxonomy".
Zachary Dankowicz
I don't see why the aforementioned "kleptotaxonomist" wouldn't be allowed to perform the alleged actions, by the Code or any other regulation. Descriptions based on images have been done (see https://zookeys.pensoft.net/articles.php?id=6143), but are quite unpopular (see https://biotaxa.org/Zootaxa/article/view/zootaxa.4196.3.9).
That said, the effort needed to use AI to find new species images alone seems like a waste of time on anyone's part. I can't imagine that anyone with an actual interest in describing new species would go to that extent. Also, while the bar of "formally publishing" something is lower in some regions, it still requires some effort and I don't expect anyone to be doing this in the near future. If they do, I wouldn't expect to wait long for a series of synonymy papers to be published in succession.
Does anyone have real examples of this occurring before internet science?
This is all possibly way too unlikely, but I'm curious to know if there are real, past examples of "kleptotaxonomy". --
You received this message because you are subscribed to the Google Groups "Taxacom" group.
To unsubscribe from this group and stop receiving emails from it, send an email to taxacom+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/taxacom/5a78a8aa-7b2f-4689-903f-8f35b9ae0119n%40googlegroups.com.
Roderic D. M. Page
Frederick W. Schueler
> Does anyone have real examples of this occurring before internet science?
typified by a stolen specimen, and this has a wikipedia entry -
https://en.wikipedia.org/wiki/Kleptotype - though with only one example
in which it's not clear that the specimen was stolen, or just a paratype
in a different museum.
fred.
------------------------------------------------------------
---------Frederick W. Schueler & Aleta Karstad ------------
Fragile Inheritance Natural History - https://fragileinheritance.ca/
6 St-Lawrence Street Bishops Mills, RR#2 Oxford Station, Ontario K0G 1T0
on the Smiths Falls Limestone Plain 44.87156° N 75.70095° W
------------------------------------------------------------
Salticidologist
PO Box 487
Morgan SA 5320
0402109743
--
You received this message because you are subscribed to the Google Groups "Taxacom" group.
To unsubscribe from this group and stop receiving emails from it, send an email to taxacom+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/taxacom/cb20528b-5840-402a-972c-37c8b987401d%40istar.ca.
Douglas Yanega
(1) in a very strict and technical sense, if someone takes a photograph of a preserved type specimen and includes it in a Code-compliant publication, then there is no violation of the Code if the specimen is lost or destroyed before the paper describing it is published, even if the author includes no statement regarding the type depository. After all, Art. 16.4.2 allows for non-extant types (the statement "where the holotypes or syntypes are extant specimens..." means that holotypes or syntypes might NOT be extant specimens). Yes, all that remains of the type is the image, but that does not, in and of itself, make the name unavailable. It doesn't even mean that a neotype is needed, if the taxon can be recognized and diagnosed without access to a physical specimen (this is why, even though so many types were tragically lost in the fire in Rio several years ago, there is no need to designate neotypes for all of them).
There is no significant difference between this scenario and an author who takes a photo and does not preserve the specimen photographed, for whatever reason.
In both cases, somewhere between discovery and publication, the type was lost. This is undesirable, but not a violation.
(2) AI can create images of species that do not exist, and it can write descriptions of species that do (or do not) exist. These capabilities have all been demonstrated before. The question we need to ask is whether we should accept this sort of thing or not. I am honestly unclear on the benefits, and fairly clear on the drawbacks, and therefore inclined to reject the use of AI specifically in the production of species descriptions. If you want to use AI to generate keys, or other parts of the text, maybe those could be allowed, but my feeling is that the description should be written by a human being, even if only because the Code has no provision for granting authorship to an automaton. It would likely still be acceptable under the present Code, but I don't think that we should, and would instead suggest that the next Code includes a few limited but necessary restrictions.
(3) Until and unless we implement a system of publication that includes a robust and mandatory review process, we will always have a small percentage of people who will act to promote their own self-interest even when it conflicts with the collective interests of the community. The perceived need for such a system tends rather directly to correlate with the level of disruption such individuals create for others in their discipline; for some of us, we never have to contend with this, for others, it is an ongoing and unavoidable mess.
Peace,
-- Doug Yanega Dept. of Entomology Entomology Research Museum Univ. of California, Riverside, CA 92521-0314 phone: 951-827-4315 FaceBook: Doug Yanega (disclaimer: opinions are mine, not UCR's) https://faculty.ucr.edu/~heraty/yanega.html "There are some enterprises in which a careful disorderliness is the true method" - Herman Melville, Moby Dick, Chap. 82