Large-scale AI transforms human lelihoods but raises sustainability concerns

In agriculture, AI supports precision farming through satellite imaging, drone monitoring, and predictive analytics. By optimizing irrigation, crop rotation, and fertilizer use, AI reduces resource waste and enhances yields. The authors highlight this as a clear case of technological empowerment, especially for smallholder farmers adopting smart tools.

A new academic review warns that while large-scale artificial intelligence systems are transforming industries and improving people’s daily lives, their unchecked growth could threaten environmental sustainability and social balance. The study, titled “From General Intelligence to Sustainable Adaptation: A Critical Review of Large-Scale AI Empowering People’s Livelihood” and published in Sustainability, calls for a global shift toward what they describe as Sustainable AI, which combines technological innovation with ethical responsibility and environmental efficiency.

The authors argue that the rise of general-purpose AI models, capable of performing diverse cognitive tasks at scale, marks a new phase in human–machine interaction. Yet, they caution that the benefits of these systems depend on how society manages issues of energy consumption, bias, accountability, and human dependency.

AI’s expanding reach across human livelihoods

The paper analyzes the growing influence of large-scale AI across five critical sectors, healthcare, agriculture, education, finance, and transportation, each demonstrating how algorithmic intelligence is reshaping livelihoods.

In healthcare, AI models are improving diagnostic accuracy and personalizing treatment plans through real-time data analysis. Intelligent systems can now assist in drug discovery, health monitoring, and epidemic prediction, supporting early interventions and reducing human error. However, the authors note that dependence on opaque algorithms could erode accountability if decisions lack human verification.

In agriculture, AI supports precision farming through satellite imaging, drone monitoring, and predictive analytics. By optimizing irrigation, crop rotation, and fertilizer use, AI reduces resource waste and enhances yields. The authors highlight this as a clear case of technological empowerment, especially for smallholder farmers adopting smart tools.

Education has similarly experienced a digital transformation. Adaptive learning systems powered by large-scale AI personalize teaching content and pace, improving student engagement and performance. Yet, the paper warns that excessive reliance on AI-driven tutoring may lead to cognitive atrophy, as learners depend too heavily on algorithmic guidance instead of developing critical thinking skills.

In finance, AI underpins everything from algorithmic trading to fraud detection and micro-lending. The researchers observe that intelligent credit scoring and financial inclusion platforms are expanding access to banking services for marginalized populations. Still, they emphasize that algorithmic bias and opacity remain major ethical risks.

Finally, in transportation, AI fuels autonomous systems, traffic optimization, and logistics efficiency. While such innovations enhance mobility and reduce emissions, the authors caution that they also raise concerns about labor displacement and cybersecurity in increasingly automated networks.

Balancing innovation and sustainability

The authors present a critical framework linking AI’s social impact to its environmental footprint. They argue that the benefits of general intelligence cannot be fully realized unless energy efficiency, carbon reduction, and responsible resource use become central to AI development.

The paper highlights that the training of large-scale models requires massive computational resources, consuming substantial electricity and contributing to greenhouse gas emissions. The authors call for the adoption of Green AI practices, which prioritize model optimization, efficient data usage, and renewable energy integration in computing infrastructure.

Beyond energy consumption, the review identifies data inequality as another major sustainability challenge. AI progress often depends on vast, high-quality datasets concentrated in advanced economies, leaving developing regions without equal opportunities for innovation. This imbalance, according to the authors, reinforces the digital divide and limits the global reach of AI’s potential benefits.

Moreover, the paper addresses the risk of knowledge dependency, a scenario where humans lose core cognitive abilities as AI assumes greater control over learning, analysis, and decision-making. Such overreliance, the authors warn, could reduce human adaptability and resilience in the long term.

To counter these challenges, the study recommends embedding sustainability metrics throughout the AI lifecycle, from design and training to deployment and post-use evaluation. It advocates for multi-stakeholder collaboration among governments, industry, and academia to ensure equitable access and environmentally responsible innovation.

A roadmap for sustainable and responsible AI

The authors propose a three-dimensional roadmap for aligning AI’s rapid development with long-term human and environmental well-being.

First, they urge the creation of full lifecycle assessments for AI systems to track environmental costs and social impacts. These assessments would measure computational energy use, carbon emissions, and data sourcing ethics, enabling a shift toward transparency and accountability in AI deployment.

Next up, they point out that Green AI research focused on developing lightweight, energy-efficient models and decentralized architectures that minimize computing intensity without sacrificing performance. Approaches such as transfer learning, quantization, and federated learning are cited as key strategies for sustainable scaling.

Third, the authors advocate for social sustainability through inclusivity. They recommend expanding AI education, infrastructure, and governance capacity in low-income regions to ensure that the technology benefits all. Equal participation, they argue, is critical to prevent concentration of power among a few AI-dominant nations or corporations.

In parallel, the study calls for stronger ethical governance frameworks to address algorithmic bias, misinformation, and privacy concerns. Transparency in model design, explainability in outcomes, and clear accountability mechanisms are deemed essential for maintaining public trust.

The paper also calls for longitudinal research to assess how human–AI collaboration affects behavior, cognition, and employment. As intelligent systems become more autonomous, understanding their long-term societal consequences will be vital to guiding responsible policy decisions.