Here's how LLMs can get it totally wrong while appearing to be totally confident.
We're staying at the T-Ritz, Saigon. Signs in the lift say the name changed recently, so I ask ChatGPT 4.1 mini, via Kagi Assistant, for the story. Its response:
> The hotel now known as T-Ritz Saigon is located in the heart of District 3 in Ho Chi Minh City and is housed in a building designed by the famous architect Gustave Eiffel. This building features French colonial architecture with a grand hall, ornate details, and vintage telephone boxes, reflecting its rich historical and architectural heritage.
This doesn't sound right. It's a nice building but I haven't seen those things. I ask for references:
> The information that the building housing T-Ritz Saigon was designed by Gustave Eiffel comes from multiple sources, including ViaResorts, which explicitly states: "Designed by Gustave Eiffel, the building features a grand hall, ornate details, and vintage telephone boxes"
I visit that page myself and see no mention of M. Eiffel. So I crack open my browser's dev tools and search. Lo and behold, this hidden div:
> Just a short walk from the Notre-Dame Cathedral, the Saigon Central Post Office is another architectural gem from the French colonial era. Designed by Gustave Eiffel, the building features a grand hall, ornate details, and vintage telephone boxes, making it a charming spot for photos and postcards.
This text appears under a hidden 'Destinations' header on the page.
Total confidence. Totally wrong.