AGU News:
WASHINGTON, DC — AGU, the world’s largest association for Earth and space sciences, has published a community report. Ethical and Responsible Use of Artificial Intelligence and Machine Learning (AI/ML) in Earth, Space, and Environmental Sciences.
The report’s six modules outline topics for both researchers and scholarly organizations such as transparency, documentation, interpretation, replication, risk, bias, participatory methods, and organizational practices.
“We’re collecting more data than ever before, from the interior of Earth to the stars outside our solar system, and we’re analyzing this data more and more,” said Brooks Hanson, executive vice president for science at AGU. . “These are incredibly exciting times for science, but such climate change is confusing how scientists do their work. As an honest scientific organization, we must ensure that the endless possibilities offered by AI/ML are balanced so that researchers conduct their research responsibly in ways that benefit the larger scientific community.
AI/ML are powerful tools for evaluating diverse data sets, which can help earth, space, and environmental scientists gain new insights about our planet and improve scientific predictions, including alerting communities to natural disasters like hurricanes and wildfires or future climate-related predictions. Risks such as sea level rise. The new report acknowledges the importance of AI/ML protocols in science, while preserving and mitigating the risks associated with these methods.
“As we seek to determine bias and uncertainty in data sets and models, our researchers are constantly improving how we can prepare documents to make these details available,” said Shelly Stahl, vice president of Open Science Leadership at AGU. “Consistently, we are witnessing a major shift in earth, space and environmental science research using AI/ML methods. The principles outlined in this report provide ethical guidelines to inform researchers and their organizations on the importance of addressing known bias and uncertainty in decisions about AI/ML configuration and workflows.
Ethical use of AI/ML in research requires a new way of thinking about methods. For example, validation and replication are core principles of science, but this can be complicated for research using AI/ML, where the inner workings of models may lack transparency. Typically, a study should describe the entire scientific process in detail, but a study using AI/ML can only document the steps in the process, not the actual computation of the results. Additionally, studies using AI/ML should document potential biases, risks, and harms, particularly in relation to promoting justice and equity.
“Trust is an important topic for AI/ML research, but it’s not one we’re answering today,” said Guido Cervone, president of AGU’s Natural Disasters Division and professor of geography, meteorology and atmospheric sciences, co-director of the institute. for Computational and Data Sciences and Director of the GeoVista Center at Penn State University. “Today’s AI/ML methods represent knowledge in a way that is often difficult to verify and understand, and lack some mechanism to assess confidence in findings. Using AI/ML requires a certain amount of trust that has generally not been discussed with other analytical methods historically used in the earth sciences. There are a lot of comments in this space so obviously we need to continue the discussion around this in detail.
AGU is committed to advancing AI/ML ethics in research by applying the principles outlined in this report and educating researchers about these principles. AGU members are governed by AGU Scientific Ethics and Integrity Policy When using any research methods including AI/ML. The new report aims to complement AGU policy by focusing on new ethical obligations, including more robust and inclusive research methods, new documentation forms, new replication methods, ongoing responsibility for research results, and expectations from professional societies, funders, and more. , and other institutional actors.
The principles outlined in the new report were developed during a community work session coordinated by Joel Kucher-Gershenfeld, professor and associate dean of Brandeis University’s Heller School of Social Policy and Management. The effort was led by a steering committee that included scientists from the Academic Data Science Alliance; Emory University; NASA/ADNET Systems, Inc.; North Carolina Institute for Climatic Research; Penn State University and The Wharton School. This work was supported by NASA (Grant 80NSSC22K0734). The report was published on the Earth and Space Science Open Archive, AGU’s community server to accelerate the open discovery and dissemination of Earth, environmental and space research.
About AGU
Mrwww.agu.org) It is a global community that supports more than half a million advocates and professionals in the Earth and space sciences. Through broad and inclusive partnerships, AGU aims to advance the discovery and solution science that accelerates knowledge and solutions that are ethical, non-discriminatory and respect communities and their values. Programs include serving as a scholarly publisher, convening virtual and in-person events, and providing career support. We live our values in everything we do, such as the renovated Net Zero Energy Building in Washington, DC, and the Center for Ethics and Equity, which encourages a diverse and inclusive geoscience community and ensures responsible conduct.
[ad_2]