AI is powering sustainability efforts, but its environmental impact is a critical consideration

AI can drive impact across the value chain, if used responsibly

AI is powering sustainability efforts, but its environmental impact is a critical consideration

AI can drive impact across the value chain, if used responsibly

The rise of AI has given organizations a new tool for managing climate risk, improving resource efficiency, discovering climate solutions, automating reporting, and tracking compliance.

Organizations are leaning in – 64% of businesses report using AI to advance their sustainability agenda. However, the use of Gen AI in this context has decreased: 52% of organizations surveyed reported using Gen AI for sustainability, down from 65% in 2024 and 56% in 2023. One reason for this reduction in the use of Gen AI could be concerns related to its environmental impact: 57% of executives acknowledge that this is being discussed in boardrooms. And, while agentic AI promises wide-ranging applications, the technology is still too early-stage to influence organizations' climate strategies.

Organizations have reason to be cautious in implementing AI, given the technology’s potentially large environmental footprint. While 57% of executives currently believe the benefits of Gen AI outweigh its environmental impact – this is a significant drop from last year’s 67%. In assessing AI’s impact, startup Mistral AI found that model training and inference accounted for 85.5% of their emissions and 91% of water use.⁴

Embracing AI therefore requires a balance of caution and pragmatism.

57%

of organizations are discussing Gen AI’s environmental impact – but only 32% have taken steps to mitigate it

_{⁴ Mistral AI, “Our contribution to a global environmental standard for AI,” July 2025.}

Tomás Zaborowski, Head of Sustainability Excellence at Bayer Crop Science:

There’s a huge opportunity in using AI for reporting and prediction. With the massive volume of data and the need for transparency, AI can assist in adapting, generating, and validating content against regulations and legal frameworks. At the same time, its predictive capabilities offer the potential to forecast environmental impacts or supply chain disruptions before they happen – that’s where I see real efficiency gains.”