Leveraging AI-derived Data for Carbon Accounting: Information Extraction from Alternative Sources

Olamide Oladeji,Seyed Shahabeddin Mousavi

2023-11-27

Abstract:Carbon accounting is a fundamental building block in our global path to emissions reduction and decarbonization, yet many challenges exist in achieving reliable and trusted carbon accounting measures. We motivate that carbon accounting not only needs to be more data-driven, but also more methodologically sound. We discuss the need for alternative, more diverse data sources that can play a significant role on our path to trusted carbon accounting procedures and elaborate on not only why, but how Artificial Intelligence (AI) in general and Natural Language Processing (NLP) in particular can unlock reasonable access to a treasure trove of alternative data sets in light of the recent advances in the field that better enable the utilization of unstructured data in this process. We present a case study of the recent developments on real-world data via an NLP-powered analysis using OpenAI's GPT API on financial and shipping data. We conclude the paper with a discussion on how these methods and approaches can be integrated into a broader framework for AI-enabled integrative carbon accounting.

Computation and Language,Artificial Intelligence,Machine Learning

What problem does this paper attempt to address?

The paper primarily explores how to leverage artificial intelligence (especially natural language processing technology, NLP) to improve the reliability and accuracy of carbon accounting. Specifically, the authors focus on how to better track carbon emissions in the supply chain by extracting information from unstructured text data (such as company earnings call transcripts and shipping documents).

The main contributions of the paper include:
1. **Proposing the use of AI technology to address carbon accounting issues**: Given the limitations of traditional methods in providing accurate, timely, and detailed carbon emission data, the paper proposes the adoption of AI technology (particularly NLP) to enhance the effectiveness and precision of carbon accounting.
2. **Demonstrating the application of NLP in supply chain tracing**: Through specific case studies, the paper demonstrates how NLP technology can be used to extract information about a company's supply chain from earnings call transcripts and construct a dynamic knowledge graph to track and manage carbon emissions.
3. **Analyzing shipping document data**: The paper also explores how publicly available shipping document data (such as bill of lading records) can be used to identify specific companies' suppliers and types of goods, thereby estimating the carbon emissions of these goods.
4. **Performance evaluation**: The paper reports on the performance of a large language model based on GPT-3 in extracting information about buyers, suppliers, and goods, showing high accuracy and good practicality.

In summary, this paper aims to demonstrate how NLP technology can effectively extract key information about carbon emissions from alternative data sources, thereby supporting more accurate and comprehensive carbon accounting practices. This not only helps to improve supply chain transparency but also promotes the achievement of global emission reduction goals.