Universities and businesses are embracing AI at breakneck speed while systematically ignoring the environmental cost of the computing power that runs it. A growing body of research reveals a troubling gap: institutions that have spent years building climate-conscious reputations are now deploying enterprise-level AI tools without measuring or accounting for their energy consumption and carbon emissions. The result is a fundamental contradiction between stated sustainability commitments and actual operational practices.

Why Are Organizations Ignoring AI's Environmental Impact?

The problem starts with a convenient accounting loophole. When universities and corporations use cloud-based AI services, the energy consumption and emissions occur at remote data centers owned by third parties. Because these emissions happen off-campus or off-premises, many organizations exclude them from their official greenhouse gas inventories. This creates a false sense of climate progress.

At American University, for example, the institution achieved carbon neutrality in 2018 through aggressive energy efficiency and a massive solar partnership in North Carolina. Yet the Kogod School of Business has since rolled out enterprise-level access to generative AI tools for every student without communicating the environmental costs of that expansion. The university's sustainability plan does not treat digital infrastructure, such as data centers and AI tools, as its own environmental category.

The challenge is compounded by a measurement problem. Over half of the electricity powering data centers in the United States comes from fossil fuels, and these centers consume billions of gallons of water annually. Yet universities and corporations often claim they cannot measure AI's environmental impact because "measuring AI's environmental impact is still an emerging field, especially in a university context," according to a statement from American University.

What Are the Hidden Costs Organizations Are Missing?

Researchers at Loughborough University have developed a framework called Digital Decarbonisation to make visible what organizations typically ignore. The framework identifies five critical dimensions that organizations must consider when adopting AI and digital technologies:

• Costs: The financial expenses of storing, processing, and training data on cloud infrastructure, which many organizations underestimate or fail to track comprehensively.
• Carbon: The energy demand and greenhouse gas emissions generated by data centers, model training, and continuous AI operations that occur outside organizational boundaries.
• Compliance: Emerging regulatory, reporting, and governance requirements that will increasingly demand transparency about digital infrastructure and AI-related emissions.
• Security: The resilience and cybersecurity risks created by dependence on complex digital supply chains and cloud-based AI services.
• Supply Chain: The dependencies and vulnerabilities embedded in global networks of cloud providers, hardware manufacturers, software vendors, and data infrastructure.

The research emerged from a strategic discussion convened at the House of Lords on April 27, 2026, bringing together senior expertise from technology companies, standards organizations, defense agencies, and academic institutions including AWS, IBM, and IEEE.

Digital waste represents a particularly acute problem. Organizations are generating, storing, and processing more data than ever before. Duplicated data, inefficient systems, poorly governed storage, unnecessary processing, and unmanaged AI use create avoidable costs that ripple across energy demand, carbon emissions, cyber resilience, and supply-chain dependency.

How Can Organizations Balance AI Growth With Climate Responsibility?

The tension between AI adoption and environmental responsibility is not insurmountable, but it requires deliberate action. Experts argue that sustainability should not be framed as a brake on digital or AI growth. Instead, handled well, Digital Decarbonisation can support growth by helping organizations become more efficient, more resilient, and better prepared for future regulatory and infrastructure pressures.

For businesses, this could mean reducing unnecessary cloud and data costs, improving digital governance, and making stronger investment decisions. For small and medium-sized enterprises (SMEs), it could mean practical guidance on adopting AI without creating avoidable cost, risk, or complexity. For public services, it could mean more responsible use of AI, data, and digital infrastructure.

The broader challenge is that AI has brought the issue into sharper focus. Applied AI depends on data, compute power, cloud infrastructure, skilled users, energy systems, hardware, and global supply chains. As AI tools become embedded in everyday work, education, public services, and business operations, organizations will need to understand not only what AI can do, but what it demands.

Data centers powering AI technologies are consuming what researchers describe as "terrifying" amounts of energy. A recent analysis from London revealed that the data centers powering these technologies are consuming electricity at levels that rival, and sometimes exceed, those of entire nations. Training and operating massive AI models requires super computing powers running 24/7, making the data center sector one of the fastest-growing energy consumers globally.

While tech giants have begun investing heavily in wind and solar energy and developing innovative cooling systems, experts argue these efforts are slow compared to the speed of AI growth. Observers believe that balancing digital progress with environmental preservation has become an urgent necessity that cannot be delayed.

Steps Organizations Can Take to Address AI's Hidden Environmental Costs

• Conduct a Digital Audit: Map all cloud services, AI tools, data storage, and digital infrastructure in use across the organization to identify where energy consumption and emissions are occurring, even if they happen off-premises.
• Set Measurable AI Sustainability Goals: Establish concrete targets for AI-related energy use, data center emissions, and carbon footprint of model training, rather than treating AI as exempt from climate commitments.
• Implement Digital Governance Standards: Create policies that guide when and how AI tools should be used, including decision frameworks for when computationally intensive tools may not be environmentally preferable to alternatives.
• Engage Cloud Providers on Emissions Transparency: Require vendors to report the carbon intensity of their data centers and the energy consumption associated with specific AI services used by the organization.
• Integrate Digital Infrastructure Into Climate Plans: Treat data centers, cloud computing, and AI tools as explicit categories within organizational sustainability strategies, not as invisible externalities.

The House of Lords discussion identified four practical areas where Digital Decarbonisation can support responsible AI adoption. These include SME-specific AI guidance to help smaller organizations understand the implications of AI use; reframing sustainability as an enabler of AI growth rather than a constraint; developing AI regulation and behavioral change strategies that shape how people use AI and what data they store; and creating AI education and skills training that prepares leaders to understand the hidden impacts of digital activity.

American University's experience illustrates the stakes. The institution has made its name on climate progress, achieving carbon neutrality through years of deliberate effort. Yet Kogod's rush to become an AI-first institution, offering paid enterprise-level access to generative tools and saturating its curriculum with AI-based courses, has proceeded without communicating the environmental costs of that shift. The result is a clear dissonance between the university's climate reputation and its operational reality.

The opportunity ahead is significant. Organizations can position themselves not only as leaders in AI adoption, but as leaders in responsible, efficient, and sustainable digital growth. This requires making the hidden environmental, financial, and governance impacts of digital activity visible, measurable, and manageable. Without that transparency, the promise of AI-driven innovation will come at a cost that climate commitments cannot afford to ignore.