February 25th, 2025

2 min read

The promise of artificial intelligence (AI) is vast, from boosting productivity to revolutionizing industries to solving complex societal challenges. The way we implement AI today will shape the future of our societies, economies, and environments. 

International bodies such as the G7, OECD, and Global Partnership on Artificial Intelligence (GPAI) have made significant strides in establishing AI governance frameworks. These frameworks focus on core principles such as fairness, sustainability, privacy, and security. But how do we turn these policies into actionable industry practices? How can businesses build the infrastructure required to take advantage of the power of AI while also ensuring that its deployment is responsible and sustainable? 

This topic was explored in depth at the recent International AI Policy: Outlook for 2025 conference hosted by the Center for Strategic and International Studies (CSIS), at which NTT DATA Inc. CEO Abhijit Dubey was a keynote speaker.
Watch his presentation here. 

The Challenge of Infrastructure: Scaling AI Responsibly 

AI’s potential to drive economic growth is staggering. A recent report from McKinsey examined 63 use cases and estimated that generative AI could add the equivalent of $2.6 trillion to $4.4 trillion annually. AI is set to radically transform industries ranging from healthcare and education to manufacturing and finance.  

However, the rapid growth of AI is not without its challenges, and one of the most pressing is the burden on infrastructure. AI technologies, particularly those driven by machine learning models, require enormous amounts of computational power. Data centers, which are the backbone of AI deployment, are expected to see a 160 percent increase in power demand by 2030, according to a report from Goldman Sachs. This surge will be accompanied by a significant rise in carbon dioxide emissions and a growing strain on water resources used for cooling these data centers. 

At the same time, trust in AI is dwindling. A recent survey from Edelman found that global trust in AI companies has fallen from 61 percent to just 53 percent in the past five years. This highlights a critical challenge for businesses: How can they make the most of AI in a trustworthy, transparent, and environmentally sustainable way? 

The Pillars of Responsible AI Deployment 

To navigate these challenges, businesses must focus on three key areas:  

  • Sustainability

    The energy demands of AI-driven data centers and the strain they place on resources like water for cooling are unsustainable in their current form. Companies need to innovate not just for performance but for sustainability. At NTT, for example, we are investing in renewable energy and advanced cooling technologies to mitigate the environmental impact of our data centers. Additionally, we are developing energy-efficient AI models like tsuzumi, a lightweight AI model that requires a fraction of the energy compared to traditional large-scale models. We are also exploring the potential of next-generation networks, such as IOWN (Innovative Optical and Wireless Network), which uses photonics instead of electronics to reduce energy consumption in data transmission and computing.

  • Ethics

    AI must be developed and deployed in a way that upholds fairness and transparency, and is aligned with human values and societal needs. For businesses, this means ensuring that algorithms are not only free of bias but also explainable and accessible. At NTT, we have developed an AI Charter, grounded in six principles that promote an ethical approach to AI, including sustainable development, human autonomy, fairness and openness, security, privacy, and communication and co-creation.

  • Privacy and Security 

    As AI systems become increasingly integrated into business operations, protection of sensitive data takes on ever greater importance. Businesses must ensure robust security measures to safeguard AI systems against threats throughout their lifecycle. Additionally, AI developers must adhere to strict privacy protocols to protect user data and build public trust. NTT’s commitment to responsible AI, for instance, includes a zero-trust security model, where all systems are designed to operate under the assumption that no user or system is inherently trustworthy. This approach ensures that security is maintained across every stage of AI deployment, from development to operation.

A Call to Action 

AI’s potential is immense, but it can only be fully realized if we also address its environmental, ethical, and security challenges. The infrastructure required to support AI must be sustainable, and the governance frameworks that guide its development must prioritize fairness, privacy, and transparency. 

By embracing a responsible approach to AI, through sustainable practices, ethical governance, and global collaboration, we can ensure that AI delivers on its promise without compromising the future of our planet or society. This is the future we must build together.