Sustainability assessment using multimodal artificial intelligence agents

Оценка устойчивости с использованием мультимодальных агентов искусственного интеллекта
Zhihan Zhang, Vikram Iyer, Gregory D. Abowd, Shwetak Patel, Adriana Schulz, Tingyu Cheng, Zachary Englhardt, Alexander Le Metzger, Yuxuan Mei, Felix Hähnlein
2026-06-12

electronic device carbon footprintemission factor predictionlife-cycle assessmentlife-cycle inventorymultimodal multi-agent AI
Reducing the growing environmental impact of the computing industry requires assessing the emissions of electronics at scale. However, a traditional life-cycle assessment (LCA) of an electronic device, which maps materials and processes to environmental impacts, often requires proprietary or unavailable data. Here we report a multimodal multi-agent artificial intelligence system that emulates the collaborative process between LCA professionals and stakeholders (such as product managers and engineers) to estimate the carbon footprint of electronic devices. The agents iteratively construct a complete life-cycle inventory by leveraging a structured data abstraction and software tools that mine information from the public Internet, including repair communities and government regulatory databases. This reduces data gaps and data collection from weeks or months of expert time to under 1 min. The system can calculate the carbon footprint within 19% of expert LCAs with zero proprietary data (typical of the variation between human LCAs). We also show that by encoding domain-specific knowledge, environmental impact estimation can be reframed as a data-driven prediction task, in which both unknown products and emission factors are represented as weighted combinations of similar ones with known emissions. A multimodal multi-agent artificial intelligence system that emulates the collaborative process between life-cycle assessment professionals and stakeholders (such as product managers and engineers) can calculate the carbon footprint of electronic devices, performing data collection tasks that can take weeks or months of expert time to under a minute.
1
A multimodal multi-agent AI system estimates electronic-device carbon footprints by emulating collaboration between LCA professionals and product stakeholders.
2
Automated data collection reduces tasks that typically require weeks or months of expert effort to under one minute, without using proprietary data.
3
Domain-specific knowledge reframes environmental-impact estimation as data-driven prediction using weighted combinations of similar products and emission factors with known impacts.
4
The system estimates carbon footprints within 19% of expert LCAs, matching the typical variation observed between human LCAs.
5
The system iteratively builds complete life-cycle inventories using structured data abstractions and tools mining public Internet sources, including repair communities and government databases.

electronic devices

life-cycle carbon-footprint estimation, including automated construction of life-cycle inventories from public data and prediction of unknown products and emission factors

Publication Details
Publication Date
2026-06-12
Journal
Publisher
ISSN
Cited by
12
Access Type
Author Information
Authors
Zhihan Zhang
Vikram Iyer
Gregory D. Abowd
Shwetak Patel
Adriana Schulz
Tingyu Cheng
Zachary Englhardt
Alexander Le Metzger
Yuxuan Mei
Felix Hähnlein
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%