AI must not become the newest form of resource extraction

The global artificial-intelligence buildout is entering a decisive phase. Data centres can bring lower-income countries investment, better connectivity, tax revenue and technical capability. But without strong safeguards, the same projects could reproduce an old economic pattern: foreign companies capture the profits while local communities supply cheap land, electricity and water, then inherit the environmental and financial costs.

The risk is not hypothetical. The International Energy Agency projects that global electricity demand from data centres will more than double by 2030. The United Nations Environment Programme has also warned that many countries still lack comprehensive frameworks for sustainable data-centre operations. As AI expands, data-centre policy is no longer a niche technological concern. It is a central question of infrastructure, public finance and economic development.

The concern is not that developing countries should reject data centres. Digital infrastructure can support hospitals, schools, government services and domestic businesses. The concern is unequal bargaining power. A government seeking investment may offer discounted land, long tax holidays or subsidized electricity without fully calculating the public cost. A headline investment can look impressive even if it creates relatively few permanent jobs or requires taxpayers to finance grid and water upgrades.

The first safeguard should be radical transparency. Before approval, governments should publish expected electricity and water demand, proposed subsidies, tax concessions, infrastructure requirements and projected permanent employment. Operators should report actual resource use and performance against their commitments. Public resources should not be negotiated through secret contracts.

Second, essential community needs must have legal priority. A data centre should not receive discounted or guaranteed power while households and small businesses endure unreliable service or rising costs. The same principle applies to water. In stressed watersheds, project reviews should evaluate competing demand from residents, agriculture and industry, as well as lower-water cooling alternatives.

Third, incentives should be earned, not gifted. Tax relief should be tied to measurable public benefits such as locally financed grid capacity, renewable generation, wastewater reuse, technical apprenticeships, university partnerships and domestic supplier development. If promised outcomes are not delivered, benefits should be reduced or recovered. Governments should compare the value of every subsidy with taxes collected, infrastructure costs and durable employment.

Fourth, host countries need enforceable environmental standards. Measures such as power-usage effectiveness, water-usage effectiveness and renewable-energy use should become permitting and reporting requirements rather than optional corporate pledges. Independent audits, drought restrictions, e-waste plans and financial guarantees for site cleanup should be standard conditions of operation.

Fifth, countries should protect their digital sovereignty. Hosting servers does not automatically create control over the data, models or economic value running through them. Public-sector and sensitive citizen data need clear rules for location, access, encryption, cross-border transfer, incident reporting and deletion. Government contracts should also require portability and realistic exit rights so that public institutions are not trapped with one foreign cloud or AI provider.

Finally, developing countries should cooperate. If neighboring governments compete by offering weaker regulations, cheaper public electricity and longer tax holidays, multinational operators gain leverage while citizens lose. Regional minimum standards for disclosure, environmental protection and taxation could reduce that risk.

The right question is not whether data centres are good or bad. It is whether the host country receives a fair share of the value and retains the authority to protect its people. A responsible project should pass a simple test: residents keep priority access to affordable water and power; contracts and subsidies are public; environmental limits are measurable; local skills and infrastructure improve; citizen data remain protected; and cleanup is financially guaranteed.

Artificial intelligence should not become the newest form of resource extraction. With transparent contracts, enforceable community benefits and strong sovereignty protections, data centres can become genuine development infrastructure. Without them, the AI economy risks repeating a familiar history under a digital name.

Aminur Chowdhury is a senior process engineer at Canon Nanotechnology, with a PhD in materials science and engineering from UT Austin and an ASQ Certified Six Sigma Black Belt.

Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the opinions and views of The Business Standard.