AI must serve humanity, not endanger it

Artificial intelligence is unlike any technology humanity has seen before. Its awe-inspiring power is breathtaking, but it needs a sound global governance system. Other transformative technologies in the past, such as electricity, the steam engine, automobiles and thousands of other inventions, reshaped civilization without the need for international oversight.

This is because AI is different. First, it can execute many human tasks in ways never seen before, but it can also be misused to harm society through fraud, fake narratives, cybercrimes and data manipulation. Unless AI is properly regulated, the harm it can do far outweighs the benefits it offers. But if well-governed, AI could take us to the next level of human evolution.

This is why the BRICS Leaders' Statement on the Global Governance of Artificial Intelligence, issued in Rio de Janeiro in July, is timely. It rightly emphasized "the need to achieve safe, ethical, trustworthy, and responsible AI development for the benefit of all" and called for systems that can detect and prevent malicious use. These principles reflect a growing global consensus that AI must serve humanity, not endanger it.

Second, AI governance has to be global in nature because its benefits and potential misuse impact the whole world. Many countries are formulating laws and strategies for AI governance. These efforts are necessary, but not sufficient. Fragmented AI governance leads to duplication, hinders standardization and fails to stop countries trying to maintain their technological hegemony.

For example, the United States passed the CHIPS and Science Act of 2022 to block China's plans to develop high-performance AI models. Despite this, Chinese firms are making spectacular advances in AI. The US shows little commitment to the goal of "AI for All". This was evident when, along with the United Kingdom, it refused to sign the final Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet at the Paris AI Action Summit in February 2025. The declaration was signed by over 60 other nations, including China, India and France.

The third reason for global governance of AI is more fundamental. AI is racing toward Artificial General Intelligence (AGI), which promises a dramatic integration of man and machine. This integration can already be seen in the emergence of humanoids. As AI gets integrated into our biological systems and AI-to-AI communication proceeds autonomously, humanity will confront some existential questions: Will machines replace humans? What is the meaning of being human? How can humans control machines instead of being controlled by them? No single country can answer these questions.

These concerns have prompted some of the world's leading thinkers to urge caution. Geoffrey Hinton, a Nobel laureate in physics and a pioneer in deep learning, warned in 2023 of an "existential threat" posed by advanced AI systems and even called for a temporary halt in AI development.

But technological progress cannot be stopped. The global community must find practical ways to minimize the risks of AI and maximize its rewards for humanity as a whole. In this context, the words of Zhang Jun, secretary-general of the Boao Forum for Asia and China's former United Nations ambassador, are noteworthy. Speaking at the UN Security Council debate in July 2023, he stressed that AI's impact may exceed human cognitive limits and called for "people-oriented" and "AI for good" principles to regulate the development of AI and prevent the technology from "turning into a runaway wild horse". He said humanity must retain the ability to press the stop button when required. Cautioning about the abuse of AI in conflict situations, he also made a fervent appeal: "All countries should uphold a responsible defense policy, oppose the use of AI to seek military hegemony or to undermine the sovereignty and territorial integrity of other countries."

India has taken a similar approach in its AI Governance Guidelines issued in November. The country's goal is to encourage innovation and adoption, while protecting individuals and society from any harm caused by AI.

The global debate on AI governance must address three critical concerns. First, it must prevent, and reverse, the widening AI divide within and among nations. It is estimated that AI could generate $15.7 trillion in additional value by 2030. If these gains remain concentrated in wealthier nations and their tech giants, the Global South will face deeper inequality, weakened digital sovereignty and new security risks.

Second, with AI usage growing rapidly, there is an urgent need for global cooperation in energy conservation. The world cannot afford a new wave of greenhouse gas emissions driven by data centers and computing power. Ways must be found to maximize "AI per watt of energy".

Third, the question "What should AI be used for?" must be at the core of its global governance. Should AI be used for world peace or warfare? Should its priority be fulfilling the UN Sustainable Development Goals or making the wealthy even wealthier? Should AI be used to protect forests, de-pollute oceans and rivers, clean the air in cities, ensure people's food and health security and enhance the quality of education in the world or should it be used to intensify geopolitical rivalries?

At the UN headquarters in New York, there is a bronze sculpture with a timeless message: "Let us beat our swords into plowshares". Some negative forces in the world, who are clearly in the minority, may want to misuse AI as the newest and most lethal sword invented in human history. But the right-thinking majority must resolve to beat it into the best ever plowshare to grow prosperity and happiness for all.

The author served as an aide to India's former prime minister Atal Bihari Vajpayee, and is a votary of India-China friendship and cooperation.

The views don't necessarily represent those of China Daily.

If you have a specific expertise, or would like to share your thought about our stories, then send us your writings at opinion@chinadaily.com.cn, and comment@chinadaily.com.cn.