How systematic entomology will thrive in the age of artificial intelligence

By Jiri Hulcr, Ph.D., Andrew J. Johnson, Ph.D., and G. Christopher Marais

G. Christopher Marais

Andrew J. Johnson, Ph.D.

Jiri Hulcr, Ph.D.

In a recent New York Times profile, Mauricio Diazgranados, the new director of the New York Botanical Gardens, shared a message that many scientists are grappling with: “We cannot continue to do science as we are used to doing. Can we continue to work in the field, introducing and describing new species while the entire world is being destroyed and burned?”

Does Mauricio’s message resonate among entomologists? Our field, systematic entomology, often struggles with this question: should we focus on providing solutions to contemporary problems or should we continue to document the world’s biodiversity? By spending our careers documenting insect species and their relationships instead of solving the world’s immediate problems, are we potentially threatening the survival of our field?

We believe history shows the answer: documenting biodiversity, yes here to stay as one of the foundations of biological sciences. What we need is to continue to evolve As we do it. Our survival will be the result of our ability to reorganize ourselves.

Systematists have evolved repeatedly

A few decades ago, traditional systematic entomology was derided as stamp collecting, an obsolete field that would soon be replaced (then) by cladistics, phylogenetics, and studies of evolutionary processes. It turns out that systematics as a whole has survived and thrived, while cladistics not so much.

A decade or two later, molecular biology took over. Taxonomy was once again predicted to become extinct and taxonomists to be considered a dying breed. Yet as we filled GenBank with millions of DNA sequences, we found ourselves unable to place them in meaningful context, or even label them with names. Instead of dying out, experts familiar with organisms have become a hot commodity, and new National Science Foundation programs have increased funding to document biodiversity.

Systematics has survived and thrived because each new generation of systematists has embraced the new tools that time has brought.

How systematics can become “machine intelligible”

What’s hot now? Artificial intelligence (AI). Language models and image recognition could be the next disruption in biodiversity documentation. When iNaturalist is in everyone’s pocket, it’s time to evaluate what systematic entomology needs to do to stay relevant, well-funded, and thriving.

Our lab played with machine learning, which produced two results: a prototype AI bark beetle classifier and a keen awareness of the importance of people with in-depth knowledge of the organism. Once again, as we adapt our field to use new tools, we need the humans who crawl through bushes, sift through leaf litter, and spend time peering into museum drawers to be the arbiters of what it is biological truth versus what is an artifact. of an algorithm. And, once again, taxonomists may become increasingly important, but only to the extent that we are able to cooperate with the machine.

This is what it means for systematists to be “machine intelligible.”

1. Our results must satisfy the data hunger of machines. Even the good old entomology of “stamp collecting,” like the accumulation of specimens, has become valuable again. The key will be to turn samples into data and make them available. So, to our fellow systematists: please post your images and label them generously. Post your morphological descriptions in extensive detail. Publish field observations and host associations. Machines continue to collect our data from the web; make sure yours are there. If your collection isn’t online, it doesn’t help the common cause. If we give the models enough morphological terms, one day you will be able to identify your insect by talking about it to your computer.

2. We humans should be more disciplined about our vocabularies. This doesn’t mean we have to write like robots. While the recent era of relational databases required strict consistency in format and spelling, the new era of natural language models, fuzzy matching, and graph databases does not strictly require it. What is most important is the volume and repetition of accurate statements. Statements need not be unified or even grammatically correct; rather, they must be factually accurate. From now on we will have to be much more careful to distinguish what we know for sure and what is a hypothesis. If we are unsure, it is our responsibility to express doubt.

The taxon we study, the bark and ambrosia beetles, is a good example. Thousands of publications report the death of trees by these beetles. So if you ask ChatGPT if, for example, cockroaches in the Ips generally kill trees, he will report with great certainty that they do so. But this is not true for the vast majority Ips species, including almost all in the United States. This response is the result of an overemphasis in published work on a European tree-killing pest, while at the same time we collectors and systematists consistently fail to report when everyone else Ips beetles are only secondary colonizers of dead trees.

One more word about the language: Maybe you’re already using the Darwin Core format for your data; Great. But now let’s think about it a little differently. Don’t think of language rules as restrictions. Instead, think about the need to tell everything you know, even if it is repeated and boring. Discover ontologies and try adopting one in your work.

3. We need to keep collecting! Even as AI models get smarter and smarter, ultimately it will be you, your specimens, and your knowledge of them that will guide the machines into the murky waters of truth and knowledge. AI routinely distorts human beliefs about the world simply by sampling biases, and it doesn’t know it. With the global homogenization of biota on the one hand and rampant extinctions on the other, generating real data, not simulations, has become more important than ever. Here in Florida, the state Department of Agriculture’s Division of Plant Industries, an agency heavily involved in applied science, is developing a regional taxonomic center staffed by human taxonomists, not machines. Why? Because their biggest problem is recognizing and documenting new invasive parasites that no one has ever seen before.

The future of describing the natural world

Exciting times lie ahead for systematics. What happens after AI machines are trained for systematic work? What will come after the point where machines are more accurate than people at predicting the identity of organisms? What happens when the processes to train them have also been automated? How will the taxonomy survive? At the moment we are still describing, sequencing and photographing the biological world, but robots could soon be better at these things and take on much of the hard work. What role will expert taxonomists play then? Where will the frontier of the field be located, the tasks for which we can still “outperform” machines? We do not know yet.

We know that taxonomists are poised to fill a translational role between machines and people, both among other scientists and the public. We may need more training to interpret the complexities of the world in digestible ways. In other words, we may be needed to bridge the gap between machines, people and nature itself.

Jiri Hulcr, Ph.D.is an associate professor in the School of Forestry, Fisheries, and Geomatic Sciences and the Department of Entomology and Nematology at the University of Florida and a principal investigator in the UF Forest Entomology Lab. Email: hulcr@ufl.edu. Andrew J. Johnson, Ph.D. (ajj@ufl.edu) is a research assistant and G. Christopher Marais (gmarais@ufl.edu) is a master’s student and graduate assistant in the laboratory.

Related