+91-9958 726825
Beyond the veil of technology: The Socio-Ecological costs of AI in Business
|
Artificial Intelligence (AI) is a newly emerging technology that has transformed society due to its vast range of use cases. It has especially had a large impact on the way businesses are operated, from digital marketing, customer service through chatbots, inventory management, etc. AI is being utilized in almost every business activity. In the current landscape of rapid adoption of Artificial Intelligence (AI), it is thus essential to critically evaluate whether this technology ultimately yields benefits to society as a whole. Although numerous studies have been conducted on the positive impact of AI, its overwhelming social and ecological costs have been overlooked. The current literature has the overarching themes of the impact of AI on Sustainable Development Goals (SDG), AI Ethics, as well as AI’s positive impact on sustainability. This paper thus explores the direct ecological costs, i.e., the carbon footprint of AI, as well as the indirect social costs, including the marginalization of many minority communities. The paper utilizes methodological triangulation to highlight such socio-ecological costs, illustrating how AI, as a technology, still has areas where it can be optimized, especially in business practices that use society’s vast resources for profit. Companies and businesses must therefore be made aware of such costs in order to ensure greater accountability and corporate governance. In the conclusion, this paper aims to present the future scope of AI-related studies in India as well as practices that can be implemented on the corporate level to reduce such socio-ecological costs
|
|
There was once a time when Artificial Intelligence (AI) was predicted to become the fifth generation of the computer. The dawn of AI came sooner than expected; however, with the onset of AI Agents such as DeepBlue and Apple’s Siri. In the simplest of words, AI refers to a method or technique or ‘algorithm’ that allows machines to simulate human intelligence in order to create something new or solve problems. In the words of Marvin Minky, one of the founders of AI, AI is “enabling machines to do things that require human intelligence”[1] .
In the present day, Artificial intelligence (AI) is causing a revolution in business and society by changing the existing relationship and therefore interactions between stakeholders and individuals[2] . The prevailing outlook among companies regarding AI is extremely positive. After all, with machines and automation taking over business operations, companies can look forward to reduced costs, increased efficiency, and rising profit margins. Given the massive impact that AI has had, changing the way in which society as a whole thinks, functions, and interacts, it is necessary to know the socio-ecological costs of AI in business. Berger, S. states that social and ecological costs are interrelated concepts in existing capital markets and thus refers to socio-ecological costs as “preventable resource depletion, environmental disruption, and social problems, such as unmet human needs”[3] . The use of AI in business implies the setting up of data centers for energy, water for cooling the hardware, and large storage demands. It is assumed that training AI alone consumes up to 1,287 megawatt hours of electricity (enough for 120 average U.S. homes for a year) and generates about 552 tons of carbon dioxide. This does not include the environmental costs of dirty mining procedures and the use of toxic chemicals in obtaining raw materials for fabricating GPUs. The cool water used in the process also has massive environmental impacts[4] . The human input required by Large Language Models during the training phase is typically outsourced to independent contractors in low-income countries. These workers face downright toxic and exploitative working conditions. LLMs are also being used for spreading fake news, propaganda, data theft, and bias. If no action is taken soon, AI has the potential to cause harm to our biosphere and undermine social and democratic processes[5] . The attitude of critical optimism that businesses and other organizations alike have adopted is inadequate in tackling this complex technology. The first step in preventing this is thus to equip both businesses and those affected by AI with the necessary knowledge to ensure the ethical and sustainable utilization and development of Artificial Intelligence[6] . Thus, the study attempts to ascertain the socio-economic costs of AI and identify measures and areas for policy intervention. The study further adopts an Information Systems perspective through the Socio-Technical Systems Theory and IT Governance Theory. These frameworks help explain how AI in business is not merely a technical tool but rather an interconnected socio-organizational system that influences environmental sustainability, social welfare, and governance mechanisms. Despite the increasing discourse on AI ethics and sustainability, existing studies focus either on energy consumption and carbon emissions or on ethical concerns such as privacy and bias. There is limited research that examines the socio-ecological costs of AI in business operations through an integrated perspective combining environmental degradation, labour exploitation, governance failures, and sustainability concerns. This creates a significant gap in understanding how the use of AI in businesses affects both ecological sustainability and social welfare simultaneously. The study aims to address this gap by critically evaluating AI as a driver of ecological and social costs through an integrated perspective. It also proposes policy frameworks to support sustainable AI in business. |
Research Questions RQ1: What are the major ecological costs associated with the use of AI in business operations?
RQ2: What social risks and inequalities emerge from AI business systems? RQ3: What policy frameworks can support sustainable AI in business? |
|
This study adopts a qualitative conceptual research design based on the analysis of the literature and policy suggestions. Methodological triangulation has been used to collect the requisite data. Primary and secondary data in the form of company BRSR, Sustainability Reports, research papers, and telephonic interviews with five environmentalists have been utilized. The selection of five environmentalists was guided by the principles of methodological triangulation and qualitative depth rather than statistical representation. These participants were chosen for their expertise and engagement with environmental issues with the aim to eliminate bias and provide in-depth understanding of the complex theme. Interviewees were selected based on a snowball sample, and their insights were divergent from some of the information collected through company reports.
The collected data were analyzed using thematic analysis. Themes relating to ecological costs, social costs, governance gaps, and sustainability interventions were identified using an interpretation of the literature and interview responses. The Google Scholar and SSRN databases have been used to cover most articles. Keywords such as “AI and sustainability”, “socio-ecological cost of AI”, “environmental cost of AI data centers”, “AI ethics”, and “AI and SDG” have been employed to search for articles in the databases. |
|
The existing literature puts more emphasis on analyzing or computing AI-related energy use and greenhouse gas (GHG) emissions[7] [8] [9] . While these papers contribute to the discourse on the ecological costs of AI, they fail to consider the broader impact of using AI in business practices on environmental degradation, including aspects of water consumption, resource degradation, waste generation, as well as the indirect and adverse social effects. Additionally, some papers focus only on specific subsets of AI, such as Large Language Models (LLMs) or Generative AI (GAI). While studies that focus on AI’s socio-ecological costs holistically do not consider the context of business and organizational settings[10] [11] .
Some key social issues related to AI as entailed by existing papers include the lack of data privacy, algorithmic discrimination and bias, and job loss [12] [13] . However, these issues are often isolated from environmental concerns, leading to two separate streams of research. As a result, there is limited understanding of how the social and ecological consequences of AI interact and reinforce one another within present business environments. This largely curtails the development of a comprehensive assessment of AI's social impacts. Further, much of the research portrays AI as a catalyst for socio-economic sustainability rather than as a driver of much of the harm. AI is given credence simply for its potential to improve efficiency, optimize resource use, strengthen environmental sustainability effort, and support the achievement of the Sustainable Development Goals (SDGs)[14] [15] [16] [17] . Nishant[14] for example, maintain that the true value of AI lies in how it facilitates environmental regulation and sustainability. Similarly, several studies assess the impact of AI in helping accomplish SDGs and suggest measures to ensure a positive shift of AI in supporting the attainment of these goals[15] [16] [17] . In contrast, research by Christensen E. and Jallat F. & Morillon G contend such optimistic outlook by emphasizing material and social costs associated with AI systems, including increased resource consumption, environmental degradation and labour displacement[10] [11] . The paradoxical perspectives remain unresolved because most studies evaluate either the benefits or the costs of AI but in isolation. Thus, the net impact of AI in business practices on sustainability continues to be unclear. In terms of policy frameworks for countering the adverse effects of AI on the environment, there exists a similar contrast. Studies in this area emphasize the need for norms, regulations, and guidelines to uphold moral values and human rights in the process of deploying AI in business practices[18] . Ethical guidelines issued by private companies, research institutes, and government organizations aim to ensure ‘ethical AI,’ while regulations aim to mitigate risks such as data security[19] [20] . However, there is yet to be a consensus on the effectiveness of such governance frameworks. Munn[21] argues that principles so put forth are often difficult to apply as they are isolated and lack, as businesses tend to largely ignore ethics and focus on profits. Therefore, it is entirely up for debate whether AI ethics and governance regulations in the current discourses are capable in mitigating the social and environmental costs of AI in any meaningful sense. Taken together, the literature reveals three key limitations that constrain the comprehensive view of socio-ecological costs of AI in business. First, the research on environmental aspects of AI focus on energy consumption and carbon emissions while neglecting broader ecological impacts such as water use, resource depletion, and waste generation. Second, social and environmental costs are typically examined separately, thereby limiting the understanding of their connection and subsequent effects. Third, studies on governance and ethics concentrate on theoretical guidelines and not on policy mechanisms capable of addressing the socio-ecological impacts. This study addresses these gaps in the existing literature in three ways. First, it develops an integrated socio-ecological perspective on AI in business, combining ecological and social dimensions. Second, the paper critically evaluates AI as a driver of ecological degradation and social inequality rather than as a sustainability enabler. Third, the study proposes a policy-oriented framework for India, where AI adoption is rapidly increasing despite limited regulatory preparedness. |
|
Ecological Costs
The adoption of Artificial Intelligence in business operations have enhanced their efficiency, automated routine processes, and improved decision-making capabilities. However, these benefits are accompanied by ecological costs, often overlooked in discussions surrounding the technology. Most of these drawbacks arise due to something called an AI data center and GPUs (Graphics Processing Units). An AI data center is an energy-consuming and power-hungry computing, processing and storage infrastructure for the training and deployment of AI services[22] . On the other hand, GPUs enable the training and deployment of AI models by processing large amounts of data[23] . Together, these components allow businesses to use AI in their operations and lead to serious ecological harm. Energy Consumption and Emissions Water Consumption Resource Consumption and E-Waste Generation Social Costs Theoretical Analysis IT Governance Theory Policies and Managerial Implications Policy Implications
The most important step is to ensure a comprehensive law that directly regulates AI and mandates the disclosure of key information, such as AI-related job loss, greenhouse gas emissions, and waste generation. Providing an ethical code of conduct and codified laws similar to the EU AI Act regarding compliance can help keep companies in check and prevent any exploitation. These laws can further prevent the spread of misinformation and propaganda through AI tools.
The Government can mandate the setting up of mentorship and allyship programs in listed companies with AI operations to help lower-level employees upskill and reskill in AI skills and prevent massive job displacements. It is also necessary to ensure the inclusion of employees engaged in MSMEs, especially in the unorganized sector, given the fact that these enterprises are the backbone of the Indian economy. This can be done through the help of not-for-profit organizations, as well as free government courses and vocational training.
An important policy intervention for developing economies like India is to mandate the phasing out of thermal/ fossil fuel-based energy to greener sources within a few years of setting up, as well as investing in alternative ways of operating data centers, including orbital data centers that cut down the water-stress on local areas[45] .
Managers can follow the examples of NVIDIA and Google to set up systems and technologies that make them more resource-efficient and their data centers sustainable. NVIDIA[46] mentions the use of 100% renewable energy sources, closed-loop cooling systems, and responsible sourcing of minerals to counter the damage done by AI data centers. Large corporations can also look into developing and adopting more energy-efficient algorithms and models.
Given the rapid spread of AI in business operations, it is necessary to address data privacy issues efficiently in order to prevent the loss of customer trust. Data shuffling refers to shuffling values of confidential attributes. It is a data masking technique that ensures data security and protects the privacy of users[47] [48] . Managers can employ other masking techniques, such as encryption, as well, in order to ensure ethical use of customer data and retain customer loyalty.
Businesses can leverage AI solutions themselves to counter their socio-ecological costs. Supply chain optimization, intelligent energy grids for energy efficiency, AI-powered robots and applications for employee safety and electronic vehicles (EVs) for logistics and transportation are some examples of the use of AI for sustainability[49] . AI can thus be leveraged to find realistic and innovative solutions to such social and environmental harms.
Table 1 : Summary of Ecological and Social Costs[25] [30] [32] [34] [35]
|
|
As observed, the use of Artificial Intelligence in business has high socio-ecological costs. The study attempts to identify these costs and suggest policies to alleviate these harms. However, the study has key limitations which may provide guidance for future researchers-
Future studies conducted on these areas can expand the understanding of the actual effects of AI on the environment and society and cover the gaps in the existing literature. |
|
The paper looks at environmental costs such as GHG emissions, energy use, water withdrawal, and unsustainable methods of sourcing resources. It also notes social costs such as data privacy issues, job loss, exploitation of workers, and imbalanced access to the benefits of AI. Lastly, it attempts to present certain policy frameworks to counter these socio-ecological costs. Any new technology or idea brings with it an inherent fear and uncertainty regarding its use and the changes it may trigger, and the same can be observed for Artificial Intelligence (AI). However, bringing in the required legislation and measures to control AI and alleviate its negative impact can ensure that this technology brings about positive changes in society and the world at large.
|
Pain Text:
Ruchita Nandy, Dr. Aruna Dev Rroy (2026), Beyond the veil of technology: The Socio-Ecological costs of AI in Business. Samvakti Journal of Research in Business Management, 7(1) 52 - 68.
Ruchita Nandy, Dr. Aruna Dev Rroy (2026), Beyond the veil of technology: The Socio-Ecological costs of AI in Business. Samvakti Journal of Research in Business Management, 7(1) 52 - 68.
