What’s that critter? New tech guidelines can help ensure we get the right answer
In a recent study published in the Proceedings for the National Academy of Sciences, biologist Julie Allen and an international working group developed a set of recommendations to ensure trustworthy biodiversity monitoring data.
Have you ever been in the woods and wondered what bug you’re hearing or which bird is singing? These days, new technologies are helping both scientists and the public identify what’s trilling, scuttling, tweeting, rustling, or slithering around a given area.
But not all emerging tech emerges equally, which is why biologist Julie Allen and an international working group developed a set of recommendations to ensure trustworthy biodiversity monitoring data. Their nine suggestions were published March 4 in the Proceedings of the National Academy of Sciences.
Allen’s credentials for biodiversity surveying are iron clad. In 2024, her team earned the grand prize in XPRIZE Rainforest for autonomously surveying 100 hectares of tropical rainforest in 24 hours and producing the most impactful real-time insights within 48 hours.
The XPRIZE highlighted the increasingly critical role of biodiversity monitoring with the number of plant and animal species declining in many parts of the world. The worst-case scenario? Species are gone before they can even be identified.
“Governments, businesses, and policymakers need to be able to make fast decisions about how to manage land,” Allen said. “To do that, they need to understand the biodiversity of a given area, and that data doesn't really exist for a lot of places.”
Good news: A whole suite of emerging technologies can be combined to help scientists rapidly assess biodiversity. For example, remote sensing via satellite images can be used to understand land changes; artificial intelligence (AI) models can identify birds, frogs, bats, and insects through image and sound; and environmental DNA can provide a snapshot of all the organisms in a given place from DNA collected from substrates like water, soil and air left behind from animals.
Bad news: “There are some real challenges. The data is all over the place,” said Allen. “It's collected differently and adheres to different standards, and for many places does not exist.”
According to Allen, guardrails will help ensure that the data collected is trustworthy.
With this goal in mind, Allen spent three intensive days in Washington, D.C., with a group of other leaders in biodiversity surveying, which included conservation groups, policymakers, bioinformaticians, computer scientists, technicians, representatives from Indigenous peoples, and private companies.
Their resultant recommendations address the growing needs of data for rapid and appropriate conservation actions. These included:
- Agreeing on standard methods for data collection
- Ensuring new technologies are accurately calibrated
- Creating living databases of trusted information to reduce the risk of poisoning by AI hallucinated, or false, information
- Ensuring respectful incorporation of Indigenous Knowledge
The next step is figuring out to how to implement these recommendations, which is something Allen will be working through as she launches her own company, EarthIntell. To start, she will be establishing best practices for surveying with environmental DNA, but ultimately hopes to expand to a complete line of monitoring technologies.
“The world is changing really, really fast,” Allen said. “Our decisions about land management need to be not only fast — but also based on reliable and accurate data.”
Original study DOI: 10.1073/pnas.2519345123
