Environmental impacts of generative AI

Technological advancements are generally considered to be helpful for the environment (e.g., tracking of gas emissions, greener forms of energy; Kühner et al., Reference Kühner, Hüffmeier and Zacher2025). Although historically these advancements are congruent with environmental goals, this pattern does not extend to generative AI. Table 1 identifies elements of generative AI and the energy costs to power those elements.

Table 1. AI Elements and its Corresponding Energy Use

It is evident that the widespread adoption of generative AI has environmental consequences, as it consumes vast amounts of energy to power its training (McQuate, Reference McQuate2023), daily operations (Li et al., Reference Li, Yang, Islam and Ren2023; McQuate, Reference McQuate2023; Berthelot et al., Reference Berthelot, Caron, Jay and Lefèvre2024), and outputs (McQuate, Reference McQuate2023; Todorović, Reference Todorović2024). As a result of the environmental costs associated with generative AI maintenance and usage, this commentary argues that employees’ use of generative AI should be considered a CSB.

Generative AI usage as a counterproductive sustainability behavior

Companies that value environmental sustainability commonly implement green policies and practices (e.g., training and development aimed at increasing sustainability; Benevene & Buonomo, Reference Benevene and Buonomo2020). The collection of policies and organizational practices that integrate the environment into human resource management is considered green human resource management (GHRM; Faisal, Reference Faisal2023).

Once hired, an employee who appreciates green policies may engage in employee green behaviors, or EGBs (Katz et al., Reference Katz, Rauvola, Rudolph and Zacher2022). EGBs are an integral part of a company’s environmental sustainability effort and are considered the core of environmental behavior within organizations (Zacher et al., Reference Zacher, Rudolph and Katz2023). Meta-analyses have found that EGBs also contribute to positive organizational attitudes such as organizational identification, organizational commitment, and job satisfaction (Katz et al., Reference Katz, Rauvola, Rudolph and Zacher2022). Therefore, EGBs are beneficial to organizations, beyond the organization’s stance on environmental policies.

Conversely, CSBs are conceptualized as an extension of counterproductive work behaviors that encompass the domain of sustainability (Dilchert, Reference Dilchert2018). CSBs include employee behaviors that detract from organizations’ environmental goals and/or waste natural resources. For an organization with a focus on sustainability, employees engaging in CSBs could result in negative consequences, such as increased costs to sustainable production and/or damage to that organization’s green reputation (Ahmad et al., Reference Ahmad, Jamali and Khattak2025). Although these consequences have not yet been directly studied in the context of generative AI use, organizations that encourage over-reliance on large language models are, in a sense, encouraging employees to engage in CSBs.

This is largely driven by the understanding that CSBs are defined by their detrimental effects on environmental sustainability. Example items to measure CSB may include behaviors such as “I stick to typical ways of doing work, even though it is environmentally unfriendly; I waste resources” (Dilchert, Reference Dilchert2018, p.52). Because generative AI usage at volume consumes a vast amount of natural resources (Berthelot et al., Reference Berthelot, Caron, Jay and Lefèvre2024; Li et al., Reference Li, Yang, Islam and Ren2023; McQuate, Reference McQuate2023; Tordorović, Reference Todorović2024), this commentary takes the stance that, by conceptual and operational definition, the use of generative AI within the workplace falls under the category of CSB.

Employee generative AI considerations for practice

This commentary acknowledges that employee use of generative AI at work is becoming ubiquitous, with over 50% of the workforce utilizing it to help perform their jobs (Brooks, Reference Brooks2023). As organizations rapidly adopt these tools, often in pursuit of efficiency or competitive advantage, they may unintentionally overlook the environmental costs associated with AI deployment. Although these impacts are increasingly visible in popular media, they are not always factored into internal decision-making. Rather than offering prescriptive policy solutions, we suggest organizations and industrial-organizational (I-O) psychologists engage with a series of guiding questions that promote reflection on sustainable AI practices.

Are organizations overlooking sustainability costs in everyday AI use?

Although generative AI offers powerful workplace efficiencies, it is also a resource-intensive technology. The environmental burden of training and using large language models—measured in energy use, carbon emissions, and water consumption—is well-documented (McQuate, Reference McQuate2023; Todorović, Reference Todorović2024; Li et al., Reference Li, Yang, Islam and Ren2023). Still, these costs are often diffuse, downstream, or framed as infrastructure issues rather than behavioral ones. How might organizations evaluate whether their current AI usage patterns align with sustainability goals? How can organizations make the environmental costs of workplace AI more visible and assess whether current usage aligns with sustainability goals?

What organizational norms and training practices could support more responsible AI use?

GHRM literature suggests that training, expectations, and culture shape employee engagement in sustainability efforts (Renwick et al., Reference Renwick, Redman and Maguire2013; Dumont et al., Reference Dumont, Shen and Deng2017). Might similar approaches encourage employees to think more critically about their digital resource consumption? For example, what would it look like to incorporate the environmental cost of generative AI into onboarding, training, or job performance expectations? How can employees be empowered to see indiscriminate AI use as a form of CSB?

How can strategic alignment and governance enable more sustainable AI adoption?

Technology adoption is most effective when it serves, rather than dictates, organizational strategy (Carr, Reference Carr2003). Yet in the absence of external regulation, how can organizations ensure AI use is aligned with their stated environmental values? Would incorporating environmental impact assessments into AI project planning help prevent misaligned investments or reputational risks? Could cross-functional governance teams, including sustainability officers and IT leaders, better evaluate the long-term implications of AI adoption? How might organizations expand their environmental social and governance (ESG) frameworks to include metrics such as AI-related carbon emissions or water use?

What role can I-O psychologists play in supporting sustainable AI integration?

By posing these questions, we hope to broaden the conversation about the role of I-O psychology in the age of AI. As organizations navigate the complex trade-offs between innovation and sustainability, they’ll need more than technical solutions; they’ll need behavioral insights, systems thinking, and strategic foresight. Given the behavioral, strategic, and infrastructural dimensions of AI adoption, I-O psychologists are well-positioned to contribute to this emerging area by combining scientific understanding and practical guidance that aligns technological innovation with ecological stewardship.

The widespread use of generative AI challenges long-held assumptions that technology and sustainability naturally align. By reframing indiscriminate AI use as a CSB, organizations can surface hidden risks and make more intentional choices. Embedding environmental responsibility into the way AI is adopted and utilized requires thoughtful design of policies, practices, and culture—areas where behavioral science plays a critical role. Ultimately, the environmentally responsible use of AI will require organizations to make thoughtful decisions, balancing performance, profit, and planetary health in the evolving future of work.