If local opposition pushes AI data centers offshore, Americans’ data could fall under foreign censorship, tariffs, and instability, undermining US data sovereignty.

If artificial intelligence (AI) companies are boxed out of every jurisdiction in the United States, they may need to locate their infrastructure offshore instead. That may resolve not-in-my-backyard concerns and prevent utility rates from increasing—at least from data centers—but it presents other disadvantages that AI’s opponents are likely overlooking.

The most obvious is the economic loss caused by pushing out a booming industry. Another is losing autonomy over Americans’ data.

Why Offshore Data Centers Put Americans’ Data at Risk

Consider what happens to US data that relies on infrastructure in countries that do not share our value of freedom of speech. Our data could be subject to the same censorship laws of our host nation.

Governments in countries in the Middle East control internet infrastructure and restrict connectivity by blocking encrypted services like WhatsApp, FaceTime, and Facebook Messenger. These governments also censor content that is anti-Islamic, critical of those governments, obscene, and offensive.

The region also faces political instability that could affect data center performance and reliability. Amazon Web Services recently announced that customer data stored in data centers in Bahrain and the United Arab Emirates was unrecoverable after drone attacks by Iran in March and April.

In another scenario, a host nation may place tariffs on data transmitted back to the United States or use our data as a bargaining chip. In March, the World Trade Organization’s multilateral moratorium on customs duties for electronic transmissions lapsed. The United States-Mexico-Canada Agreement also prohibits tariffs on data transfers, but the future of this agreement is uncertain due to its ongoing review process and the recently renewed trade war between the United States and Canada.

Or say AI developers go to any of the European, South American, or East Asian countries that favor carbon emissions taxes or high fuel-use taxes. Like tariffs, taxes would be costs passed down to US consumers.

These are some of the risks the United States faces if enough jurisdictions do not allow data center development domestically. Outside its boundaries, the United States loses control. 

How AI Already Shapes Americans’ Daily Lives

Yet disillusionment with AI chatbots and disdain toward AI-generated content are growing. These objections are reasonable. People value human connection, critical thinking, and human multimedia rather than life-like imitations.

But rejection of AI is not as simple as abstaining from consumer-facing, general-purpose AI chatbots or objecting to AI-generated media. AI is embedded across systems in society, and regular people encounter uses of AI more than they realize.

For example, facial recognition technology or Face ID that allows users to unlock their phones, sign into password-protected accounts, and verify their age for age-restricted platforms with a quick scan of their face is enabled by AI software.

Prospective homeowners engage with home lending advisors who use AI-powered software to verify income and financial records and extract data from applications and other documents. AI can improve objectivity and efficiency of application reviews. Banks and credit-card companies use AI to analyze patterns and deviations in transactions for improved fraud detection.

Many physicians use AI to record and summarize clinical encounters. Modern medicine involves administrative tasks to maintain a medical history for patients, which helps inform patients and their providers. But these tasks are time-consuming and take away time providers can spend with patients. AI frees up some of providers’ time spent on administration.

Additionally, more physicians use the AI-powered medical search engine OpenEvidence to assist with complex diagnoses and research the latest treatment options. Modern medicine is advancing constantly; it is challenging even for specialists to remain informed of every new clinical trial, technique, or other development in their field. Unlike general-purpose AI tools, the platform complies with the Health Insurance Portability and Accountability Act (HIPAA) and is trained not on the open internet but on peer-reviewed medical literature, which helps ensure its output is accurate and relevant. The tool is only made available to health care professionals with verified credentials.

Drivers relying on GPS navigation obtain real-time traffic information, alerts, and faster routes due to AI. AI analyzes historical and live traffic conditions to estimate arrival times, predict slowdowns, and determine routes.

So whether or not you embrace AI, you likely unknowingly use it or you are on the receiving end of its use. And each of these examples involves processing and collecting sensitive data that should be safeguarded.

Why the United States Needs Data Centers at Home

US AI data sovereignty means Americans’ data captured by AI systems is processed and stored in the United States. It also means it is governed and protected by US laws and regulations and not subject to the whims of foreign governments. Maintaining data sovereignty requires sufficient data infrastructure.

This is why it matters that we build AI data centers in the United States. Wishing for AI’s extinction will not yield the desired results, and neither will widespread moratoriums on data center developments.

But in this current environment where data centers are politically unpopular, the future of data sovereignty is at stake.

Ratepayers are right to be concerned with how potential data center developments may affect their utility bills and access to necessary resources. Homeowners and buyers are right to be concerned with how it impacts homeownership and quality of life.

AI companies and data center developers must be honest, transparent, and fair—generous, even—to escape their villain trajectory. State and local leaders have an opportunity to respond to their constituents’ concerns and make favorable deals—ones that raise revenue for their jurisdictions and ensure that ratepayers are not suffering. Moratoriums on development and excessive demands stall, if not kill, negotiations.

If data center developers have nowhere to go domestically, they will have to go abroad. It could have similar consequences on the US economy as offshoring in other industries, and it would be a shame after all the efforts to reshore manufacturing and other supply chains. Most certainly, it would relinquish data sovereignty.

About the Author: Annie Chestnut Tutor

Annie Chestnut Tutor is a policy analyst in the Center for Technology and the Human Person at the Heritage Foundation. Previously, Annie worked in the US Senate for eight years, most recently as the Commerce, Science, and Transportation Committee staffer for Senator Ron Johnson (R-WI), and as a legislative assistant for Senator Mike Lee (R-UT), where her policy portfolio included tech and telecom. She holds a bachelor’s degree in international relations from Wheaton College. 

The post Blocking Data Centers at Home Could Put Americans’ Data at Risk Abroad appeared first on The National Interest.