The rapid advancement of generative AI (GenAI) is reshaping various industries, bringing unprecedented efficiency and innovation. Sustainability reporting has emerged as a critical area, where GenAI can enhance organizations’ ability to meet increasing disclosure demands from regulators, investors, and customers. According to a 2024 Reuters survey, 67% of sustainability professionals expect GenAI to significantly impact their reporting efforts.
In this blog, we will explore five use cases of GenAI in improving the accuracy and efficiency of environmental sustainability assessments, from automating information extraction to generating reports and co-piloting for analytical applications.
Generative AI is a class of artificial intelligence models that can create new content (text, images, audio, and video) by learning patterns from existing data. It is underpinned by deep neural networks, particularly transformer-based foundation models like the Large Language Model GPT-4. These models produce highly relevant content, making GenAI a powerful tool for communication, content creation, and information analysis.
Sustainability reporting often struggles with understanding complex, evolving frameworks, analyzing unstructured data from various sources, and customizing results for diverse stakeholders.
GenAI, especially its text-based applications, is well-suited to solve these challenges. It excels at summarizing information, processing unstructured data, generating insights, and communicating findings in human language. It is efficient and scalable, although oversight is necessary to validate its outputs.
In contrast, traditional AI (discriminative AI) primarily focuses on validating and analyzing structured data, optimizing processes, and predicting outcomes based on input data. While this article focuses on GenAI use cases, combining both GenAI and traditional AI can provide a comprehensive approach to enhance the accuracy, efficiency, and depth of sustainability reporting.
| Generative AI (GenAI) | Traditional AI/ML | |
| Functions | – Extract & summarize information – Process & analyze unstructured data – Generate insights & content – Communicate in human language | – Validate data & detect anomalies – Analyze structured data – Optimize processes/resources – Perform predictive modeling |
| Features | – Efficient – Scalable – Oversight required to validate outputs | – Reliable – Accurate – Manual effort required to compile data & interpret results |
The challenge
Sustainability reporting faces an intricate and dynamic landscape. Regulatory frameworks (like CSRD/ESRS, CA SB 253/SB 261, and SEC) coexist with voluntary frameworks (like GRI, CDP, TCFD, and TNFD), making it difficult for companies to stay informed and compliant.
How GenAI can help
GenAI can help companies understand the compliance requirements in two steps:
With this information, companies can identify compliance gaps, compare their progress against best practices, and stay ahead in their obligations.
In practice
Nasdaq’s ESG AI solution leverages a curated knowledge base of company and regulatory documents to enable compliance benchmarking and gap analysis. Its AI assistant and intelligent search deliver meaningful insights to facilitate informed, timely decisions.
AI compliance assistant by Nasdaq Sustainable Lens
The Challenge
A materiality assessment helps companies identify the most impactful ESG issues for reporting. It involves engaging stakeholders to prioritize issues and handling massive documents, such as sustainability reports, regulatory filings, stakeholder feedback, industry benchmarks, financial statements, and internal policies and procedures.
“Double materiality” requirements under CSRD further complicate the process, as companies must assess the financial impacts of ESG issues and the social/environmental impacts of the business.
How GenAI can help
GenAI can streamline this process by automatically parsing unstructured text data, evaluating sentiment, and generating insights on material impacts. It effectively facilitates stakeholder communication and enhances the efficiency and comprehensiveness of materiality assessments.
In practice
Powered by C3 Generative AI, the energy technology company Baker Hughes was able to quickly parse 3,500 stakeholder documents and 400,000 paragraphs to identify the 10% most relevant ESG topics. This automation saved 30,000 hours for their materiality assessment in a two-year cycle.
Material assessment automation by C3 AI and Baker Hughes (Source: C3 Transform 2024)
The Challenge
Credible, granular sustainability reporting relies on analytical methods like life cycle assessment (LCA) and advanced software systems. It requires collecting and managing data from a wide range of sources (internal systems, supply chains, third-party databases) in various formats.
This data challenge especially hinders progress on Scope 3 emissions, which can account for up to 90% of a company’s total carbon emissions. Only 41% of companies disclosed Scope 3 emissions in their 2022 CDP report, due to limited data access, low quality, and a lack of data extraction tools.
How GenAI can help
GenAI can transform data from various sources into a common database schema (a structured framework defining how data is organized and related) to ensure accurate data mapping for analysis. It achieves this by comparing source schemas to the target schema, identifying conceptually equivalent fields despite different labels, and converting data to consistent formats and structures.
Combined with traditional machine learning, GenAI can automate tedious tasks like data validation, anomaly detection, cleaning, and normalization to improve data quality and ingestion efficiency.
In practice
Aligned Incentives’ GenAI-powered system AITrack enables companies to quickly assess the granular environmental footprint of each product across their corporate portfolio—built on custom, process-based LCAs at scale. The system accepts various data types, including Bills of Materials (BOMs), supplier data, and 3rd party lifecycle inventories. It effectively evaluates data accuracy and harmonizes data across hundreds of sources for accurate LCA modeling and reporting.
Speak to our team to learn more 🡲
Data ingestion automation by Aligned Incentives
The Challenge
Generating comprehensive sustainability reports, as required by CSRD, CDP, and investors, is a time-consuming process. It involves synthesizing vast amounts of information and tailoring the report format and language to meet the diverse needs of stakeholders.
How GenAI can help
GenAI can automatically integrate information from various data sources and write clear, relevant corporate report sections aligned with regulatory standards and stakeholder requirements, ready for manual review. Additionally, GenAI can support the generation of Environmental Product Declaration (EPD) background reports by documenting product-specific LCA data, methodologies, and environmental impacts, ensuring transparency and compliance.
In practice
The GenAI solution Alan, combined with sustainability management software leadity, can generate CSRD reports automatically based on companies’ ESG data, supporting documents, action plans, and existing CRM and ERP data.
CSRD report generation by Alan and leadity
The challenge
Life Cycle Assessment (LCA) is a cornerstone for in-depth sustainability reporting. However, it requires significant data aggregation and complex system modeling. The resulting technical outputs can often be complicated for stakeholders to understand and utilize. Real-time, customized assistance can help streamline these tasks.
How GenAI can help
GenAI could be integrated into LCA software as a co-pilot, similar to Microsoft Office Copilot. It could offer interactive support throughout the process—from data management and inventory development to impact assessment and scenario analysis. GenAI could also create customized summaries and visualizations, making LCA results accessible and actionable for all stakeholders.
While this use case is not currently available, the potential for such integration holds promise for the future of LCAs.
Carbon footprint
GenAI models carry a significant carbon footprint. Training alone for a model like GPT-3 has substantial energy costs, emitting 552 tCO2e—equivalent to the annual GHG emissions of 120 cars. Their constant operation in data centers, often powered by fossil fuels, further intensifies the environmental outcomes.
To minimize the impacts, companies can use more optimized algorithms, efficient hardware, and renewable energy sources. Additionally, it is crucial to prioritize adapting pre-trained models for new tasks, rather than training entirely new ones. Fine-tuning a smaller “student” model, as demonstrated by Accenture, can achieve accuracy similar to that of a larger model with 2.7 times less energy consumption.
Security
Using GenAI in enterprise sustainability reporting raises critical security concerns. Integrating GenAI with sensitive company data, such as proprietary environmental metrics and financial information, could lead to economic losses and reputational damage if not adequately secured. These models are also vulnerable to manipulation by attackers, potentially creating misleading reports.
To mitigate these risks, enterprises must implement robust security measures, such as regular security audits, data encryption, and strict access controls.
Aligned Incentives offers an AI-powered enterprise sustainability planning solution trusted by the world’s largest organizations. Powered by granular product lifecycle assessments at scale, our platform enables companies to efficiently measure, report, and reduce their environmental footprints across extensive value chains. Our solution accelerates the path for businesses to achieve their sustainability goals while empowering them to boost sales with product-level metrics.
