Panel 3

Panel 3: Safety, Ethics, and Regulation of AI convened experts to delve into the ethical challenges and regulatory requirements surrounding AI in environmental protection and management. The forum examined the influence of AI technologies on the formulation and implementation of environmental policies and explored ways to ensure that these technologies align with ethical principles and legal frameworks. The discussion was moderated by Jialuo Zhang, Chief Representative of the Green AI Institute's New York Chapter.

Phuong Ho, a Postdoctoral Associate at MIT CEEPR, emphasized that while AI has tremendous potential to support environmental initiatives, it is essential to ensure that it is applied ethically. Ho highlighted that AI’s predictive capabilities are invaluable in environmental protection, such as in pollution control and climate forecasting. However, Ho raised concerns about transparency and the accountability of AI models, especially when they impact policy decisions that affect public welfare and environmental resources.

Yanli He, Co-founder and President of the Post-Pandemic Leadership Research Institute, discussed the need for AI to be fair and inclusive, particularly in the context of environmental justice. He stressed that AI-driven solutions must consider marginalized communities that are often disproportionately affected by environmental issues. To this end, He argued for rigorous ethical guidelines and diverse stakeholder engagement to prevent AI technologies from inadvertently perpetuating inequalities.

Jason Mohaghegh, Associate Professor of Comparative Literature and Leader of The Generator's AI Ethics & Society Research Group at Babson College, explored the philosophical dimensions of AI ethics in sustainability. He suggested that while AI can accelerate environmental protection, its rapid development brings risks related to ethical erosion if left unchecked. Mohaghegh called for an interdisciplinary approach to AI regulation, combining insights from technology, philosophy, and ethics to foster a more holistic understanding of AI’s societal impact.

Lastly, Ahmad Antar, Founder of Digital Emissions and an Instructor in Sustainable IT Studies at Harvard University, joined online to address the technical aspects of AI’s environmental footprint. Antar pointed out that the computational resources needed for AI contribute to digital emissions, which can undermine environmental goals if not managed carefully. He advocated for sustainable practices within AI development, including optimizing energy consumption and promoting transparency in AI-related emissions reporting.

In closing, the panelists collectively highlighted the importance of transparency, fairness, and accountability as central pillars for the ethical deployment of AI in environmental management. They emphasized the need for regulatory frameworks that not only support innovation but also safeguard the environment and vulnerable communities.