AI Tradeoffs: From Efficiency Wins to Ethical Questions—Path to Responsible Intelligence
Every time you ask an AI to draft an email, solve a problem, or generate a new image, you’re tapping into a vast, digital infrastructure. Human ingenuity, authenticity, and copyright debates aside, many of us have paused to consider the energy cost and environmental impacts that extend beyond our screens. Last week, Google released a technical report suggesting real progress. A single text prompt to its Gemini model uses just 0.24 watt-hours of electricity—about the same as running a standard microwave for one second, and a 33x drop in energy use and a 44x drop in carbon emissions over just the past year.
This marks a long-waited moment—one of the first transparent estimates from a major tech company, and one that points towards progress and efficiency. But do these numbers reflect the true reality of the rest of the industry? With public estimates highly variable, outside of Google, the “true” energy and water footprint of most AI models still remain a black box. The hope is that as technology improves, efficiency will follow—much as it did during the rise of the internet. Yet, as AI’s energy appetite grows, can we catch up with the environmental costs? As AI becomes ever more integrated into our lives, it’s crucial to understand the socio-environmental and cultural implications.
The AI Boom: Promise, Power, and a Hidden Bill
Artificial Intelligence is everywhere—from chatbots to logistics, live-streaming to healthcare, and gig work to education. Its impact is visible in how we work, learn, shop, and connect. A McKinsey & Company report shows the proportion of companies using AI has increased from 20% in 2017 to 78% in 2025. As AI adoption accelerates, total data center energy use continues to rise. Data centers have more than doubled their electricity use since 2017, now accounting for 4.4% of all US electricity, with projections as high as 22% by 2028 if trends continue.
A McKinsey & Company report shows the proportion of companies using AI has increased from 20% in 2017 to 78% in 2025.
OpenAI and President Donald Trump’s Stargate initiative aim to spend $500 billion to build as many as 10 data centers. Apple plans to invest $500 billion in manufacturing and data centers in the US over the next four years. The emissions footprint is equally striking, especially as much of this electricity still coming from fossil fuels. Tech giants are exploring nuclear power to meet surging demand, with Meta and Microsoft working to create new nuclear power plants to support carbon-free energy for their AI and data center needs. While the efficiency of a single prompt is improving, the big-picture energy and environmental cost of a global AI boom is still unfolding.
The AI tradeoff: how we can ensure its value outweighs its costs. Image designed by Freepik.
The question is no longer whether AI is “worth it”—but how we can ensure its value outweighs its costs.
But at What Cost?
Rising Energy Demand: While Google’s Gemini shows substantial progress, most AI inference systems globally remain opaque in their true energy and resource footprints. Training and running AI models is energy-intensive. Take OpenAI’s GPT-4 for example: it is estimated to have consumed 50 gigawatt-hours of energy—the equivalent of powering San Francisco for three days. Inference (the act of generating responses or making predictions in real time) now accounts for the bulk of AI’s energy use.
Emissions and Inequality: Much of the electricity powering AI comes from carbon-heavy grids. As data centers proliferate, local energy costs and emissions may rise, sometimes shifting the burden onto the public. By 2028 AI could consume as much electricity as 22% of all US households annually if unchecked.
Ethical Risks: Algorithmic management can drive productivity, but also raises concerns about fairness, transparency, and worker autonomy. Automation can widen gaps for those lacking digital skills or living in under-connected regions. The rapid pace of AI-generated content also poses new challenges for authenticity, trust, and the spread of misinformation.
For every advance, there’s a growing environmental and social bill. The question is, can we afford it?
Is AI Worth It? The Case for Intentional, Ethical Deployment
AI is not inherently “good” or “bad”—it is a tool. The value of AI depends on how we design, govern, and apply it. AI already boosts efficiency across industries—from manufacturing to logistics, healthcare to finance. In Asia, AI-powered algorithmic management is optimising supply chains, staffing, and even food delivery. Initiatives such as India’s Pragati: AI for Impact—a partnership between Meta and The/Nudge Institute—use AI to create livelihoods for marginalised groups, improving access to education, healthcare, and empowering local entrepreneurs. Numerous other case studies show AI addressing a range of societal challenges. The gig economy, influencer economy, and new forms of remote and platform-based work are also fueled by AI—especially in Asia, where one in four Chinese workers is a gig worker and “shoppertainment” powered by AI is revolutionising e-commerce and global economy.
The true power of AI lies not in what it automates, but in how intentionally we choose to use it.
As we build and use these tools, we all have a role in tipping the balance toward broad, sustainable benefit. Before you start your next AI project, consider the following:
1. Build for Efficiency and Equity
Smarter Models, Smaller Footprint: The AI field is moving toward smaller, more task-specific models and innovative hardware that can do more with less energy. Explore ways in which you can improve efficiency—not only for greater capital gains, but a more sustainable future.
Policy and Regulation: As the World Economic Forum notes, regulatory frameworks must balance innovation and protection, ensuring fair work, privacy, and inclusion in the digital economy. As AI technology evolves, be aware of the changing guidelines and frameworks, and always consider ethical principals, such as the EU AI Act and UNESCO’s recommendations on the Ethics of AI.
2. Intentional Design for Social Good
Societal Benefits: While AI is not a silver bullet, it can contribute to the positive advancement of society—from health and hunger to environmental conservation, crisis response, security, justice, and inclusion. Does your project address the world’s most pressing issues today?
Democratise Access: Ensure your AI solutions are built with communities, focus on community participation, open infrastructure, and overcome barriers of language, literacy, and accessibility. Consider other exmaples from Microsoft AI for Good Lab and Google.org Accelerator.
Transparency and Accountability: Open-source ecosystems and ethical design are key to making AI a public good, not just a commercial asset. Companies must be transparent about energy use, data practices, and algorithmic decision-making. Be responsible for both your actions and inactions.
3. Foster Human-Centered Innovation
Augment, Don’t Replace: AI should be used to augment human expertise, not displace it. Upskill your team, make your organisation future-ready. As MIT economist David Autor observes—automation doesn’t just eliminate jobs—it changes them.
Inclusive Skill Building: With over 60% of employers in Asia worried about digital skills gaps, coordinated efforts between governments, business, and civil society are vital to ensure workers can adapt and thrive.
AI is not inherently “good” or “bad”—it is a tool.
The Global Stakes: Thriving, Not Just Surviving
The AI tradeoff is real—but so are the potential gains. If we harness AI intentionally:
Economic Opportunity: New industries, job categories, and ways of working can emerge, with added productivity growth potential estimated at $4.4 trillion—amplifying human agency and unlocking creativity.
Sustainability: Recent gains in AI efficiency—like Google’s—prove rapid progress is possible, but the pressure is on for both innovation and accountability. Technological progress must align with environmental goals, driving companies to make conscious decisions, cut energy waste, and invest in more efficient models, hardware, and renewables.
Social Good: Businesses today are expected not only to be profitable but fulfill their social responsibilities within a ESG framework—delivering positive societal and environmental impacts. AI can be a catalyst for progress, but only if it’s designed and deployed with empathy and inclusion at its core.
The real question isn’t whether AI is “worth it,” but how we make it so. That means designing AI not just for profit but for public value, energy efficiency, and transparency. Above all, every step must be conscious and intentional.
The AI revolution is here. The outcome depends on the choices we make—starting now.
