Green Workforce - The Confluence of Organizational Support, AI, and Job Complexity in Employee Green Behavior

Organizations strive to become more environmentally sustainable. This quantitative study thus investigates the impact of perceived organizational support toward the environment on employee green behavior, and the role generative AI and job complexity play in this relationship.

Artwork symbolizing the interplay between sustainability and technology, fittingly created by natural intelligence (the author of this thesis, who created the background image) collaborating with artificial intelligence (Microsoft Copilot, who created the image of the human head and leaves).


Organizations globally strive to become more environmentally sustainable, and their employees play a pivotal role in this endeavor. At the same time, generative Artificial Intelligence (AI) has conquered workplaces, requiring the workforce to adapt to multiple new developments at once. This quantitative study thus aims to investigate the impact of perceived organizational support toward the environment (POS-E) on employee green behavior (EGB), and the role organizational support for generative AI (POS-GENAI) and job complexity play in this relationship. 


To succeed in becoming greener, organizations must empower employees to support these endeavors through their own behaviors. Connected to this, studies suggest that the complexity of an employee’s job might be a contributing factor to the success of adopting green behaviors.

In addition to corporate environmental sustainability, employees are also expected to contribute to the adoption of new technologies in an organization. Generative AI in particular is changing the way we work. My study thus examines the interplay between POS-E, EGB, POS-GENAI, and job complexity, with the aim to add to the research body and provide useful insights to practitioners.


My study confirms that there is a strong link between the support organizations exhibit toward environmental causes and the degree of eco-friendliness in employees’ behaviors, which reflects the consensus in the research body. A more unexpected finding is that if an organization simultaneously drives the adoption of generative AI in the workplace, the effect of the support toward employee green behavior weakens. Finally, job complexity appears to bear no influence on the relationship between POS-E and EGB. However, this might be attributed to the low measurement strength of the job complexity construct I used in my study.

Implications for practitioners

  • Walking the talk: Organizations desiring a greener workforce must lead by example and implement tangible policies and actions fostering corporate environmental sustainability, so that employees feel sufficiently supported to follow suit and adopt eco-friendly behaviors at work.
  • Cannibalization risk: If said organizations simultaneously aim to advance the adoption of generative AI technologies at work, these efforts might negate the positive effect of supporting EGB. Hence, careful treading is required to ensure that the drive for a more tech-savvy workforce does not inhibit the drive for an eco-friendlier one.
  • Mitigation strategies: Possible strategies to account for this negative effect could be to opt for more eco-friendly technologies or to show positive examples of generative AI being used for green projects. This could inspire employees to view environmental sustainability and technology as mutually reinforcing instead of contrasting.


I applied a cross-sectional survey design focusing on employees of a multinational company in the mechanical engineering industry. Through a Qualtrics questionnaire with 38 items, which I distributed via e-mail and Microsoft Teams, I collected 272 responses. I then removed three responses during the data cleaning process, resulting in a sample size of 269. Using the statistical analysis tool SPSS, I analyzed these datasets. First, I examined the reliability of my variables through the Cronbach’s alpha values, which were all in the acceptable range. I then tested my study hypotheses by conducting linear and additive multiple moderation regressions with EGB as the dependent variable.