China India Japan Korea Southeast Asia Economy Politics
Home Economy Feature
Economy · Exclusive

How a Google AI Overview Boosted a Fake Oil Firm Linked to Ambani and Trump Jr.

How a Google AI Overview Boosted a Fake Oil Firm Linked to Ambani and Trump Jr.
Economy · 2026
Photo · Priti Sharma for Asian Examiner
By Priti Sharma Economy & Markets Editor Jul 2, 2026 4 min read

Last month, ProPublica published an investigation into America First Refining, a Texas oil refinery startup that had quietly received investment from Donald Trump Jr. The story wove together threads involving Trump administration tariff policies, sanctioned Russian oil, and the private zoo of India's billionaire Ambani family. At the center was John Calce, the refinery's CEO, whom reporters had spent weeks scrutinizing through lawsuits, property records, and corporate filings.

As the publication date approached, the team decided to check a separate company Calce had incorporated: Brownsville Energy Storage Terminals. The company's website painted a picture of a major global player: 850 employees, 28 million barrels of oil storage capacity across six hubs, from Houston to Rotterdam to Singapore. This contradicted everything the reporters had learned about Calce's struggles to fund a single U.S. project.

Suspecting something was off, the journalists searched for the company's executives. CEO Sarah Jenkins supposedly had 20 years of experience at major energy firms. Chief technology officer David Chen had built an AI-driven inventory system. Vice President for Sustainability Dr. Sofia Rossi had spearheaded a 'Future Fuels' program. None of them left any trace online. LinkedIn profiles were dead ends. The company's Texas phone numbers turned out to belong to a baklava caterer, a taxi service, and an OB-GYN office. Numbers for facilities in the Netherlands, Singapore, and China were also non-functional.

Examining the website's source code revealed a telling note: 'This feature isn’t implemented yet, but don’t worry! You can request it in your next prompt!' The domain was registered this year through Hostinger, which offers an AI website builder for $2.99 per month. The site was entirely fabricated by artificial intelligence.

Google's AI Overview Ratified the Fiction

More troubling was that Google's AI Overview—now a default feature in search results—treated the fake company as legitimate. When a reporter searched for an award the website claimed to have won, the AI Overview stated: 'Recent notable recipients include Brownsville Energy Storage Terminals, recognized for their rapid expansion in the independent oil and terminal operations sector.'

Brownsville Energy Storage Terminals is a real LLC, but everything on its website—history, job postings, diversity policy—is fictional. Google's AI had ingested this fabricated content and presented it as authoritative. A Google spokesperson responded: 'AI Overviews are rooted in our core Search ranking systems, surfacing reliable and high-quality information for the vast majority of queries. For uncommon search terms like these, there might not be high quality information published that matches the query—and we use these examples to improve our search systems.'

After ProPublica contacted Hostinger, the company took down the site. 'After receiving your inquiry, we carried out an internal review. Based on the violations identified, we suspended the website and the account behind it in line with our Terms of Service,' a spokesperson said.

This incident is part of a broader pattern. In April, The New York Times reported that Google's AI Overviews were accurate about 90% of the time, but given the scale of searches, that still meant 'tens of millions of erroneous answers every hour.' A BBC reporter tested the system by writing a fictional article claiming he was the best tech journalist at eating hot dogs; both Google's AI and ChatGPT repeated the claim. Research from 404 Media showed that AI search results can be easily manipulated using Reddit posts.

The fake website ended up as just a single paragraph in ProPublica's investigation, but the implications extend far beyond one story. As AI-generated content proliferates, the tools we rely on for verification are increasingly vulnerable to deception. For journalists covering the intersection of business, politics, and technology in the Indo-Pacific, this case serves as a warning: the line between real and fabricated is blurring, and even the most sophisticated search engines can be fooled.

More from this story

Next article · Don't miss

Plaza Accord 2.0 Talk Won't Solve Europe's China Competitiveness Problem

German Chancellor Friedrich Merz proposes a new Plaza Accord to force yuan appreciation, but experts say it won't address China's technological and industrial advantages. The original 1985 pact contributed to Japan's stagnation, and Beijing has no interest in

Read the story →
Plaza Accord 2.0 Talk Won't Solve Europe's China Competitiveness Problem