Article by Stacey Varsani
Founder & Principal Consultant, Hadouken Consulting

IBM and L’Oréal recently announced a collaboration to use AI in developing more sustainable cosmetics. By creating a custom AI model, they aim to optimize formulations, accelerate the use of renewable ingredients, and reduce energy and material waste. This partnership highlights how AI can drive environmental, social, and governance (ESG) initiatives.
However, while AI can support sustainability goals, its own environmental impact cannot be ignored. Training AI models, especially large ones, requires substantial computational power, consuming vast amounts of energy. For instance, training a single advanced AI model can generate carbon emissions equivalent to those of multiple cars over their lifetime. Even a single query to ChatGPT uses enough electricity to power a 10W LED bulb for 20 minutes. In fact, ChatGPT's total energy consumption is comparable to that of a small country like Barbados!
Most AI companies, including OpenAI (the maker of ChatGPT), do not disclose emissions data. However, Google reported a 48% increase in greenhouse gas emissions between 2019 and 2023 due to growing data center energy consumption and supply chain emissions. Microsoft revealed a similar trend in its 2023 sustainability report, citing a 29% increase in emissions since 2020, largely from expanding data center capacity to support AI workloads.
As AI evolves with more advanced audio, video, and reasoning capabilities, its energy demands—and associated environmental impacts—are expected to rise. These challenges underscore the urgency of making AI development and deployment more energy-efficient and environmentally friendly.
What Is Green AI?
Green AI focuses on reducing the carbon footprint of AI systems by optimizing their energy consumption, leveraging renewable energy sources, and minimizing resource waste. It involves a commitment to sustainability at every stage of the AI lifecycle—from development and training to deployment and monitoring.
Examples of efforts to promote Green AI include:
Net Zero Innovation Hub for Data Centers: Google, Microsoft, and other partners launched this initiative to connect European data centers with the energy and utilities sectors. Their goals include reducing energy consumption, increasing efficiency, reusing excess heat, balancing grids, exploring alternative backup energy sources, and decarbonizing building materials. Although this is not directly tied to AI model development or deployment, it supports the broader infrastructure that enables AI systems to function.
On-Chip AI Processing: The Los Alamos National Laboratory (LANL) developed algorithms that process datasets on-chip, i.e. without a computer in the loop. This approach achieved similar speed and accuracy as traditional methods while using only 2.5% of the power.
Practical Strategies for Organizations Adopting AI
To ensure AI adoption supports ESG goals rather than undermining them, organizations should integrate sustainability and ethical practices across the AI lifecycle. Practical strategies include:
Alignment of AI Initiatives with ESG Goals: Ensure AI initiatives align with the organization's broader ESG commitments, such as reducing carbon footprint, promoting diversity or enhancing transparency.
Training and Awareness: Educate staff and stakeholders about ESG implications of AI, fostering a culture of ethical AI development.
Third-Party Accountability: Hold vendors, partners, and suppliers to high ESG standards when sourcing data, tools, or infrastructure.
Sustainable Data Practices: Optimize data storage and collection to minimize energy consumption.
Energy-Efficient Models: Use techniques like model distillation and pruning to reduce energy requirements during training and deployment.
Green Infrastructure: Deploy AI on energy-efficient infrastructure, such as renewable-powered data centers or edge computing.
Data Equity: Ensure datasets are representative and inclusive of diverse demographics to prevent bias and promote fairness.
Carbon Footprint Tracking: Continuously measure and optimize the energy usage of deployed AI systems.
Responsible Decommissioning: Plan for the decommissioning of AI models and infrastructure, ensuring data is securely deleted and hardware is recycled responsibly.
Conclusion
Green AI represents a necessary shift toward aligning technological progress with environmental responsibility. By adopting energy-efficient algorithms, leveraging renewable energy, and promoting sustainable practices, companies can significantly reduce AI's environmental impact.
Adopting Green AI practices is not only an ethical responsibility but also a strategic necessity in today's landscape of heightened public scrutiny, growing political awareness, and evolving regulatory pressures. Organizations that prioritize sustainability in their AI strategies will be better positioned to thrive in a competitive and environmentally conscious market.
If your business is looking to integrate sustainability into its AI strategy, Hadouken is here to help. We specialize in guiding organizations through complex ESG challenges, offering tailored strategies to align your initiatives with broader sustainability goals. Contact us today to explore how we can empower your business to lead in the age of Green AI.
Commentaires