Developing new technologies to protect ecosystems: Planning with adaptive management

L Luz Valerie Pascal (School of Mathematical Sciences and Centre for Data Science) I Iadine Chades (The Environmental Informatics Hub, Department of Data Science and AI, Monash University) M Matthew P. Adams (School of Mathematical Sciences and Centre for Data Science) K Kate J. Helmstedt (School of Mathematical Sciences and Centre for Data Science)

Abstract

Technology development is an essential investment for policymakers to address contemporary global crises, including climate change, biodiversity loss, the energy transition, and emergent infectious diseases. However, investing limited resources in the development of new technologies is risky. The research and development process is unpredictable, with unknown timelines and outcomes. In addition, even after successful development, the effects of deploying a new technology remain uncertain. When confronted with these uncertainties, policymakers must determine how long they should allocate resources to developing new technologies. Informed decisions require anticipating possible successes and failures of both technology development and deployment, which is a challenging optimization task when managing dynamic systems, such as threatened ecological systems. Using an adaptive management approach from AI, we find a time limit new technologies should be developed for, which balances costs, benefits, and uncertainties during development and deployment. We extract clear and transparent general rules for investing in new technologies, building on an analytical approximation. Using Australia’s Great Barrier Reef as a case study, we demonstrate that the development time limit ranges between 0 to 45 y before surrendering. We also show how characteristics of an ecological system influence the optimal investment strategy. Our approach can inform the development of new technologies in multiple domains including biodiversity conservation, public health, energy production, and the technology industry more broadly.

Article Details

Volume / Issue Vol. 122, Issue 39
Published September 30, 2025
ISSN 0027-8424
Publisher National Academy of Sciences

Authors (4)

L

Luz Valerie Pascal

School of Mathematical Sciences and Centre for Data Science

I

Iadine Chades

The Environmental Informatics Hub, Department of Data Science and AI, Monash University

M

Matthew P. Adams

School of Mathematical Sciences and Centre for Data Science

K

Kate J. Helmstedt

School of Mathematical Sciences and Centre for Data Science