A ProPublica investigation has uncovered an instance in which Google's AI-powered search feature, known as AI Overview, presented information about a company called Brownsville Energy Storage Terminals as verified fact, even though the company's website was entirely generated by artificial intelligence and described a business that does not appear to exist.
The incident began when ProPublica reporters were investigating John Calce, CEO of America First Refining, a Texas oil refinery startup with ties to Donald Trump Jr. While researching a separate company Calce had incorporated, Brownsville Energy Storage Terminals, they found a website describing what appeared to be a major multinational corporation.
The website claimed the company had more than 850 employees, 28 million barrels of oil storage capacity across six global hubs, and offices from Houston to Rotterdam. However, when reporters attempted to verify this information, they could not locate any of the company's listed executives—CEO Sarah Jenkins, CTO David Chen, or Vice President for Sustainability Dr. Sofia Rossi—anywhere online.
Reporters discovered that calls to the company's Texas phone numbers connected to unrelated businesses including a Houston baklava caterer, a Dallas-area taxi service and an OB-GYN office. Numbers for facilities supposedly located in the Netherlands, Singapore and China were also non-functional.
The site was created this year through Hostinger, which offers an AI website builder for $2.99 per month. The platform's homepage states: 'Describe it, and AI builds it.' A note in the site's source code read: 'This feature isn't implemented yet, but don't worry! You can request it in your next prompt!'
What the Right Is Saying
Conservatives and free-market advocates have cautioned against regulatory overreach in response to individual incidents of technology failures. They argue that companies like Google should be permitted to refine their AI systems through iterative improvement rather than facing government mandates.
"Every new technology has growing pains," said a spokesperson for the Computer & Communications Industry Association, which counts Google among its members. "The fact that this was caught by investigative journalism demonstrates that our existing verification mechanisms—journalism and competitive scrutiny—are functioning as designed."
Senator Ted Cruz of Texas, whose state includes both companies mentioned in the investigation, has previously opposed legislation that would impose liability on platforms for content surfaced through search algorithms. His office noted that any regulatory response should await a full examination of how AI Overviews function.
What the Left Is Saying
Progressive critics have pointed to this incident as evidence that major technology platforms are not adequately prepared for the proliferation of AI-generated content. Consumer protection advocates argue that when Google surfaces false information about companies through its dominant search engine, it creates potential for financial fraud and market manipulation.
"This is exactly what we warned would happen," said a representative from the Electronic Frontier Foundation in a statement. "When you have one company controlling how billions of people access information, and that company's AI systems can be fooled into validating complete fabrications, you've created a systemic risk to public trust."
Senator Elizabeth Warren has previously proposed legislation targeting AI-generated fraud, arguing that current consumer protection laws are insufficient to address scenarios where automated systems amplify false corporate claims. Her office did not immediately respond to requests for comment on this specific incident.
What the Numbers Show
Google commands approximately 91% of global search engine market share as of January 2026, according to industry tracker StatCounter. This means the vast majority of internet users researching companies, products or services encounter Google-generated results by default.
AI Overview was rolled out broadly in 2025 and now appears for a significant portion of queries without user opt-in. The feature uses AI to synthesize information from multiple sources, presenting summaries at the top of search results.
Hostinger's AI website builder has been marketed with promises that generated sites will "appear on Google and AI search automatically." The platform's terms of service prohibit fraudulent content, but enforcement depends largely on external reporting rather than proactive monitoring.
Following ProPublica's outreach, Hostinger suspended the Brownsville Energy Storage Terminals website. A company spokesperson stated: "Based on the violations identified, we suspended the website and the account behind it in line with our Terms of Service."
The Bottom Line
The incident raises questions about how AI systems should verify information when encountering newly created websites for uncommon search terms. Google's own statement acknowledged this limitation, saying that 'for uncommon search terms like these, there might not be high quality information published that matches the query.'
Google has stated it uses such examples to improve its search systems. The company declined to specify what changes, if any, would prevent similar occurrences in the future.
Watch for potential Congressional interest in AI oversight of search platforms. Both parties have expressed concern about misinformation risks, though they differ on whether industry self-regulation or government mandates represent the appropriate solution.