Federal AI Adoption Raises Concerns: Lessons from Cloud Computing and Cybersecurity Failures

President Donald Trump and his Cabinet assert that artificial intelligence will revolutionize the nation by enhancing prosperity, efficiency, and security. This rhetoric echoes earlier administrations, including President Barack Obama’s, which promoted cloud computing during its technological revolution a decade and a half ago.

As a cybersecurity reporter at ProPublica, I have spent the past two years examining how the federal government and its IT contractors, such as Microsoft, have navigated major technological shifts—with AI now dominating the conversation. While home users, corporations, and the federal government rush to adopt AI, the administration’s urgency mirrors past technological transitions, raising questions about long-term implications.

My reporting highlights critical lessons from the federal government’s handling of past technological shifts, offering cautionary insights as policymakers push for AI adoption across federal agencies.

Lesson 1: Hidden Costs of 'Free' Upgrades

In the early 2020s, cyberattacks attributed to Russia, China, and Iran targeted U.S. federal agencies, prompting the Biden administration to seek assistance from major tech firms. Microsoft CEO Satya Nadella responded by pledging $150 million in technical services to bolster the government’s digital defenses, alongside offering a “free” security upgrade for government customers.

Fast-forward to last year, when the Trump administration announced agreements with tech companies to provide federal agencies with discounted AI tools. Agencies could access OpenAI’s ChatGPT for $1, Google’s Gemini for 47 cents, and Grok by xAI for 42 cents. The administration framed these low costs as a way to “enhance mission delivery and operational efficiency.”

However, the allure of affordability masks a critical risk: vendor lock-in. Our investigation revealed that these seemingly free upgrades were part of a profit-driven strategy. Once agencies installed the upgrades, switching to competitors became cumbersome and costly, leaving them with little choice but to pay escalating subscription fees. A former Microsoft salesperson described the initiative as “successful beyond what any of us could have imagined.”

In response to inquiries, Microsoft stated that its “sole goal during this period was to support an urgent request by the Administration to enhance the security posture of federal agencies who were continuously being targeted by sophisticated nation-state threat actors.”

Federal agencies considering discounted AI tools must weigh the short-term savings against the potential for long-term financial and operational constraints.

Lesson 2: Overpromising and Underdelivering on Security

During the cloud computing transition, federal agencies were promised enhanced security and efficiency. However, the reality often fell short. Agencies struggled with data breaches, misconfigurations, and inadequate oversight, exposing sensitive information to cyber threats. These failures underscore the risks of prioritizing speed over security in technological transitions.

As AI adoption accelerates, similar concerns emerge. The rush to integrate AI tools without robust safeguards could replicate past mistakes, leaving federal systems vulnerable to exploitation.

Lesson 3: Lack of Standardization and Accountability

The federal government’s piecemeal approach to technological transitions has often lacked standardization, leading to inefficiencies and security gaps. Without clear guidelines or accountability measures, agencies may adopt incompatible systems or overlook critical vulnerabilities.

For AI, this risk is amplified. The absence of standardized protocols for AI deployment could result in inconsistent security practices, data privacy violations, and operational inefficiencies across federal agencies.

“The federal government’s history of technological transitions reveals a pattern of overpromising and underdelivering, often at significant cost to taxpayers and security.”

As the Trump administration pushes for rapid AI adoption, these cautionary tales serve as a reminder: technological progress should not come at the expense of security, transparency, or long-term sustainability.