In 2026, as the global ESG investment scale exceeded $3 trillion, a quietly emerging 'greenwashing' undercurrent is eroding investor trust. On July 30, Aihara Intelligence, a Singapore fintech company, released its latest AI tool, throwing a bombshell into this trust crisis—through natural language processing (NLP) technology, the system can identify false environmental claims in corporate ESG reports with 92% accuracy, installing a 'magic mirror' for sustainable investment.
The 'Greenwashing' Dilemma Under the ESG Investment Boom
As climate change issues heat up, global asset managers are incorporating ESG (environmental, social, and governance) indicators into investment decisions. According to Morningstar data, global ESG funds had net inflows of $85 billion in the second quarter of 2026, a record high. However, behind the market's enthusiasm, some companies are playing 'greenwashing' tricks: describing 99% of revenue as 'green transformation,' using vague language to cover up carbon emissions, and even fabricating environmental data.
Chen Weihong, professor at the Lee Kong Chian School of Business at Singapore Management University, pointed out: 'ESG reports currently lack unified audit standards, and many companies are good at word games, such as using uncertain words like 'possible' and 'expected' to evade responsibility. Traditional manual review is inefficient, and facing thousands of pages of disclosure documents annually, regulators are often overwhelmed.'
NLP Breakthrough: From Words to Truth
The ESG verification tool released by Aihara Intelligence this time is precisely aimed at the above pain points. The tool is based on the Transformer architecture, trained on over 500,000 global corporate ESG reports, and can deeply analyze the semantic structure, logical chain, and sentiment tendency of report text. Its core functions include:
- Contradiction Detection: Cross-comparing the main text of the report with footnote data to identify inconsistent statements, such as claiming 'carbon neutrality' without mentioning carbon credit purchase details.
- Vague Language Recognition: Marking high-uncertainty words such as 'possible,' 'probably,' and 'planned,' and counting their frequency. Exceeding the threshold triggers an alert.
- Real Data Verification: Automatically associating with the company's historical emission database, satellite remote sensing images, and energy consumption records to verify the authenticity of key indicators in the report.
- Emotional Resonance Analysis: Quantifying the overall optimism of the report through sentiment scoring and comparing it with the industry average. Both excessively high or low scores may suggest hidden risks.
In an interview with the media, Lin Huimin, co-founder and CEO of Aihara Intelligence, revealed: 'We backtested the tool on 128 ESG stocks listed on the Singapore Exchange, and it successfully identified 17 companies with significant 'greenwashing' signs, of which 9 were subsequently inquired or issued correction notices by regulators.'
Industry Impact: Reshaping Investment Decisions and Regulatory Frameworks
The release of this AI tool is causing a chain reaction. Large asset management companies are the first to try it, integrating it into their ESG scoring process. Fidelity International's Head of Sustainable Investment for Asia-Pacific, Huang Jiaxin, said: 'In the past, we relied on third-party ratings, but rating agencies have data lags. Now, AI allows us to identify risk stocks at the first moment of the earnings season and dynamically adjust positions.'
At the same time, regulators have also seen new possibilities. The Monetary Authority of Singapore (MAS) mentioned in its annual report released on July 31 that it will explore using AI for on-site inspections of ESG reports. MAS officials believe that this technology can improve manual review efficiency by 300% and significantly reduce the miss rate.
For companies, this means the cost of falsification has risen sharply. An IR director of a listed company who wished to remain anonymous said: 'We are reviewing all our disclosure materials to ensure that every third-party data has verifiable sources. In front of AI, word games are no longer feasible.'
Singapore's Unique Advantages and Future Outlook
As a leading fintech center in Asia, Singapore has already taken the lead in the AI + ESG track. Aihara Intelligence is just one of many local companies, with several other startups developing supply chain ESG risk propagation models based on graph neural networks (GNN), and auxiliary systems that automatically generate GRI standard compliant reports using large language models.
He Junxian, research analyst at Da Hua Galaxy Securities, commented: 'Singapore has unique advantages—complete data infrastructure, open regulatory environment, and the largest-scale ESG disclosure database in Southeast Asia. In the next two years, AI + ESG has the potential to become Singapore's third growth pole after green bonds and carbon trading.'
However, technology is not omnipotent. Liu Hansheng, founder of environmental consulting company Terra Nexus, reminded: 'AI can only identify contradictions at the text level. For deep-level 'behavioral greenwashing' (such as actual emission reduction measures not being implemented), physical layer data is still needed. In the future, combining AI with IoT sensors and blockchain evidence may be the real ultimate solution.'
It can be predicted that with the iteration of NLP technology and enrichment of training data, the application of AI in the ESG field will become more extensive. As Lin Huimin of Aihara Intelligence said: 'Our goal is not only to let investors see through reports, but also to make companies implement environmental commitments. When AI becomes the gatekeeper of sustainable finance, 'lucid waters and lush mountains are invaluable assets' will change from a slogan to a quantifiable investment yardstick.'
