Layout by: Courtney Joy Ocampo
Published by: Jellyssa Boniza
Date Published: July 13, 2026
Time Published: 1:29 PM
"Modern problems require modern solutions." It’s a catchy phrase, but what happens when the solution creates an entirely new environmental crisis?
To keep up with the fast-paced demands of the 21st century, the Business Process Outsourcing (BPO) sector underwent a massive transformation, making Artificial Intelligence (AI) a core part of how it operates every single day. This human-machine hybrid model has successfully unlocked next-level efficiency of corporate productivity.
Before artificial intelligence entered the scene, the BPO industry relied heavily on a technology called Robotic Process Automation (RPA). According to global enterprise software leader SAP (2026), traditional RPA acts strictly as a mechanism to automate repetitive, rule-based operations—deploying software "bots" that interact with digital systems exactly like a human would. These programs function as master digital copycats, handling routine tasks like clicking interface buttons, entering data, managing files, and transferring structured information between applications. However, unlike today’s adaptive AI models, pure RPA has no actual brain. It cannot think, learn, or make decisions on its own; it simply follows a strict checklist of digital instructions.
The current evolution of the BPO sector is a shift from these rigid, thoughtless systems into a hybrid model where advanced AI gives these digital tools a brain, allowing them to dynamically solve problems alongside human professionals.
But while AI is busy solving modern business bottlenecks, it is quietly failing the planet. The inconvenient truth of our shiny new tech infrastructure is that algorithms require massive physical computing power—and that power is currently costing us millions and even billions of gallons of liquid freshwater.
ARTIFICIAL INTELLIGENCE AS AN ENABLER
AI has significantly helped different economic sectors for efficient work. In the BPO industry, AI has contributed mainly through AI-Powered customer support, enhanced analytics and decision making, and hyper-automation. It also provides the following: data entry and processing, invoice handling, chatbots, fraud detection, and compliance monitoring. These AI functions have significantly helped BPO companies.
According to the study of National Bureau of Economic Research (NBER), access to generative artificial intelligence tools at work increases productivity by 14% on average. The study also discovered that among the 5,179 customer support agents respondents, there is a 34% improvement for amateur and low-skilled workers while minimally impacting experienced workers. This shows how AI can be helpful in the BPO industry while also providing comfort to the adjusting workers.
However, the integration of AI into the workplace is not entirely defined by positive gains and seamless efficiency; its underlying issues are just as massive as its benefits. True to its name, this "artificial" technology lacks genuine emotional intelligence, making it highly unreliable when handling emotionally charged customer interactions or complex, non-routine problem-solving. While AI thrives in data-driven sectors like finance and logistics, the communication sector still demands a distinctly human touch.
This critical gap is exactly where the hybrid model steps in. By combining machine efficiency with human intellect, this approach establishes a balanced ecosystem where human workers serve as supervisors, vetting the accuracy of AI outputs before they ever reach the client. Yet, deploying this model presents its own set of operational hurdles. Training human employees to work effectively alongside AI is a delicate balancing act, especially since advanced technology can be incredibly unpredictable, leaving some workers struggling to master the intricate technical controls.
Concentrix, one of the largest BPO companies in the country, has introduced a hybrid operating model. There have been 1,530 digital agents deployed and fully utilized across partner vendors. Since then, over 100,000 payment interactions have been managed across email, voice, and SMS. These figures demonstrate how AI can enhance efficiency in BPO operations. Other than Concentrix, companies including TaskUs Philippines, Accenture Philippines, and Asticom Group also establish hybrid models for productive work.
THE CONTRADICTION: AI AS A DISRUPTOR
While the corporate sector wrestles with these workplace adjustments, an even more silent crisis is unfolding beyond the office walls: AI's hidden, devastating impact on the environment.
When OpenAI published its foundational paper "Language Models are Few-Shot Learners” mid-2020, it didn't just introduce GPT-3 to the world, it fundamentally altered our understanding of neural networks scaling. It proved that language models could generate work with human-like fluency by scaling up the work.
According to the landmark study “Making AI Less Thirsty" by environmental and computer science researchers Pengfei Li and Shaolei Ren, the water usage effectiveness (WUE) of modern data centers is heavily strained by large language models. Their calculations state that training the GPT-3 language model can directly evaporate 700,000 liters of clean freshwater.
Expanding on this environmental crisis, a data center infrastructure report by the Lawrence Berkeley National Laboratory (2024) reveals that the massive scale-up of these computing facilities creates an immense national footprint. Their research highlights that direct server cooling, combined with the water consumed during electricity generation, drains hundreds of millions of gallons of clean freshwater daily across global networks.
UNDERSTANDING HOW AI DRAINS FRESHWATER
The mere existence of Artificial Intelligence does not drain freshwater, the over usage does. Just like our smartphones, prolonged usage causes the back or its system to heat up. This is what happens to AI systems, it heats up and when that happens, cooling systems are needed.
AI's immense thirst comes down to two cooling processes: one at the source of power, and one at the source of computation. On one end, the electric grid relies on water-intensive power plants that vent water into the sky as steam during energy production. On the other end, the data centers themselves house hundreds of thousands of buzzing servers that generate intense thermal energy. To prevent overheating and maintain optimal performance, on-site cooling mechanisms must continuously run, mimicking power plants by drawing fresh water and evaporating it away into the atmosphere.
Artificial Intelligence is now an undeniable part of our everyday lives, whether it is in our work, our schools, or our homes. As technology rapidly evolves, AI has cemented itself as a crucial engine of the global economy. But if "modern problems require modern solutions," then AI itself desperately needs a modern solution to stop it from draining our freshwater reserves and damaging the planet's environment. Through collective global effort and innovative engineering, these solutions are already moving from theory into reality. Tech giants and infrastructure developers are shifting heavy, resource-intensive AI model training exclusively to nighttime hours when ambient temperatures are naturally cooler, dramatically reducing the amount of water lost to evaporation.
At the same time, new data centers are being strategically relocated to colder global climates to take advantage of "free air" cooling, while facilities in warmer regions are adopting zero-water closed-loop systems, direct-to-chip microfluidics, and non-evaporative liquid immersion technologies (like dunking servers in special non-conductive oil or circulating fluid directly over the chips to trap heat without evaporation) that continuously recycle the exact same liquid without consuming a single drop from local municipal supplies.
The true test of the 21st century will not be how fast we can scale our algorithms, but how wisely we build them. Progress is meaningless if the very systems we design to streamline our future leave our actual world running on empty; we must ensure that our pursuit of a hyper-automated future does not come at the cost of our physical survival.
REFERENCES:
[1] Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D., Wu, J., Winter, C., . . . Amodei, D. (2020). Language models are few-shot learners (arXiv:2005.14165). arXiv. https://arxiv.org/abs/2005.14165
[2] Brynjolfsson, E., Li, D., & Raymond, L. R. (2023). Generative AI at work (Working Paper No. 31161). National Bureau of Economic Research. https://www.nber.org/papers/w31161
[3] Li, P., Yang, J., Mohammed, M. A., George, S., Xu, L., Tsung, G., & Ren, S. (2023). Making AI less "thirsty": Uncovering and addressing the secret water footprint of large language models (arXiv:2304.03271). arXiv. https://arxiv.org/abs/2304.03271
[4] Concentrix. (n.d.). Collections automation for enhanced scalability. https://www.concentrix.com/en-gb/insights/case-studies/collections-automation-for-enhanced-scalability/
[5] Oracle. (2026, February 9). Closed-loop cooling in Oracle AI data centers. Oracle Sustainability Blog. https://www.oracle.com/news/announcement/blog/closed-loop-cooling-in-oracle-ai-data-centers-2026-02-09/
[6] SAP. (2026, January 26). What is Robotic Process Automation (RPA)? SAP Business Insights. https://www.sap.com/resources/what-is-rpa
[7] Lawrence Berkeley National Laboratory. (2024). The environmental footprint of data centers in the United States. U.S. Department of Energy, Energy Analysis Division. https://energyanalysis.lbl.gov/publications/environmental-footprint-data-centers
[8] Association of American Universities. (2024, January 23). AI programs consume large volumes of scarce water. Featured Research Topics. https://www.aau.edu/research-scholarship/featured-research-topics/ai-programs-consume-large-volumes-scarce-water
[9] Ren, S. (2024). AI consumes a lot of water — but why? [Video]. TED Conferences. https://embed.ted.com/talks/shaolei_ren_ai_consumes_a_lot_of_water_but_why
[10] FydiGlobal. (2026, April 12). The future of AI and human in BPO: 2026 outlook. LinkedIn. https://www.linkedin.com/pulse/future-ai-human-bpo-2026-outlook-fydiglobal-gnxte
[11] FydiGlobal. (2026, May 3). AI in BPO industry: Evolution or extinction. LinkedIn. https://www.linkedin.com/pulse/ai-bpo-industry-evolution-extinction-fydiglobal-xdbac
