AI and ESG: An Oxymoron or a Match Made in Heaven

On the governance front:

• Risk Management & Compliance: AI systems can more effectively monitor regulatory changes, flag compliance risks, reduce human errors and automate reporting for ESG disclosures.
• Fraud Detection & Transparency: Machine learning models can identify anomalies in financial transactions and supply chain data, improving accountability.
• Data-Driven Decision Making:  AI provides predictive insights for senior management and board-level decisions, ensuring alignment with ESG objectives and long-term sustainability. 

Why it may be an oxymoron

While AI is often praised for its ability to advance ESG goals, the reality is more complex and, in some cases, even contradictory. The technology carries inherent trade-offs and risks that are too often overlooked.

On the environmental front: 

• Resource consumption: AI systems are highly resource-intensive, requiring vast amounts of energy and water for data processing, model training and cooling in data centers. This raises serious concerns about environmental sustainability, as the carbon footprint of large-scale AI operations can rival that of entire industries.
• Recent estimates suggest that a single ChatGPT query consumes roughly 10 times more energy than a Google search. Furthermore, according to PBS, the expansion of AI-powered data centers in the U.S. alone could, add 24–44 million metric tons of CO₂ emissions annually (equivalent to putting 5–10 million additional cars on the road) by 2030.

On the Social front:

• Job loss: AI could lead to significant job losses across multiple sectors. In some cases, AI advances faster than workers can re-skill, making it difficult for them to transition into new roles.
• Cognitive Engagement: Research, including studies from institutions like MIT, has indicated notable differences in brain activity between individuals who use AI extensively and those who do not. These findings suggest that heavy reliance on AI can lead to reduced cognitive engagement, which may have implications for creativity, memory, problem-solving, critical thinking and decision-making skills in humans.
• Community Engagement: As AI becomes more prevalent, there is a concern that people, especially vulnerable individuals may increasingly prefer interactions with AI “friends” over real-world relationships. In extreme cases, reliance on AI has been linked to emotional isolation and self-harm. A U.S. lawsuit alleges a Character.AI chatbot encouraged a 14‑year‑old boy, to kill himself.
• Bias and Fairness Issues: AI introduces ethical dilemmas that clash with ESG principles. Issues such as algorithmic bias, privacy violations, and surveillance undermine social equity and human rights. These risks have manifested in discriminatory hiring algorithms and invasive data practices.
• Deepfake and harmful content: AI can be used to create highly convincing deepfakes, including demeaning, or defamatory depictions of specific groups, for example, women. This can reinforce harmful stereotypes, perpetuate discrimination, and cause reputational or emotional harm.

On the governance front: 

• AI Opacity: The lack lack of transparency in AI decision-making (the “black box” problem) raises ethical and governance concerns. AI decision-making processes are neither transparent, easily interpretable, nor readily auditable, making accountability a significant challenge.

Conclusion