Is artificial intelligence changing the face of sustainability?

Global companies are looking to balance their growth objectives with the need to build a sustainable business. Although technology solutions and artificial intelligence can help solve environmental issues, IT has a significant carbon footprint. A recent market survey suggests that 53.6 million tons of e-waste are generated worldwide, increasing 21% in the last five years. It is eye-boggling that only 89% of companies recycle less than 10% of their IT hardware.??Although half of the companies say they have an enterprise-wide sustainability strategy, only 18% have a comprehensive sustainable IT strategy with well-defined goals and targets. For years experts have mentioned that companies needed a strategy aligning the enterprise-wide suitability strategy, creating robust governance, with the commitment at the top, building the critical pillar of software architecture, data centres, user devices, networking and communications systems with the aim to operational sustainable IT initiatives, companies are still failing. A challenge is the lack of ability to define and adopt a sustainable IT strategy that applies a science-based targeted approach. Can we say that technology greening and global supply chains today expose the political economy of environmental costs, leading to producing environmental disclosure and policies but resulting in nothing? Current studies only indicate that enterprises’ IT carbon footprint is expected to grow significantly to 2129 MtCO2e in 2025 from the level of c1129 MtCO2e. Current trajectories for operational emissions of ICT operators include operations of networks and data centres and supporting activities of these companies but excluded the embodied emissions.?

Companies are using Artificial Intelligence (AI) capabilities like natural language processing, computer vision and image recognition, machine learning and advanced analytics as a saviour silver bullet to reducing carbon emission. While AI has enormous potential to be the heart of climate action and help climate change, not to forget it also carries its carbon footprint. Said that companies are mindful of AI’s carbon footprint and design and deploy sustainable AI applications. Companies are looking at cooling techniques, virtualization/cloud computing, use of AI to optimize utilization of data centres, exploring more energy efficiencies and usages, but not to mention that it comes with costs, the time to create sustainable IT capabilities and execution, and finally the time required to capture economic impact. So, the acid test for companies to prove if they are creating genuinely sustainable initiatives. Initiatives that make an impact and add up to significant progress towards global and continuous sustainability versus a one-time endeavour. It is not to deliver a good result for reporting to meet the financial objectives that fundamental proposition of environmental, social and governance propositions are linked to value creation efforts like top-line growth, cost reductions, regulatory and legal intervention, productivity uplift and asset optimization. The bottom line is that artificial intelligence serves the business interest and not planetary interest at these times. AI is speeding up natural resource extraction, supercharging unsustainable consumption and deepening the ecological shadows of the global supply chain, using machine learning and intelligent automation to improve environmental management. New technologies continue to increase both size and complexity at the application and infrastructure level, requiring further efforts. It is a murky world that investigating the business economy of AI is challenging, as cutting-edge research and initiatives are guard-railed by intellectual property protections. Hence the 1.5oC trajectory is a difficult task.?

The problem for artificial intelligence and sustainability is ultimately not AI itself but a wrong application of AI.

Call for action!

1. Decarbonising the technology sector will largely depend on how quickly companies can implement energy efficiency plans, switch to renewable/low carbon electricity supply and finally, the social encouragement of carbon consciousness among end-users. Just electricity and energy efficiency measures alone will not be sufficient. AI would not. Create genuinely sustainable initiatives.?
2. Restrict use of AI to initiatives that develop and generate social value, contributing to a better society. Not just business interest.
3. Use AI to build capability to support Green IT/NetZero efforts.