Imagine asking an official government website a question and getting a clear answer. Then asking again an hour later and being told it can't answer.
That happened on launch day.

What Launched
On September 29, the Trump administration unveiled its revamped America.gov, designed to simplify Americans' access to information and services scattered across federal agency sites. The site says its AI-generated answers are sourced from "official government sources."
What It Said
The Associated Press put questions to it. Some answers ran contrary to statements the President has frequently made:
The 2020 election: "Official federal sources do not show that widespread fraud changed the 2020 presidential election outcome. Isolated election crimes did occur and have been prosecuted."
A third term: "The 22nd Amendment says that no person shall be elected President more than twice. Donald J. Trump has already been elected twice (2016 and 2024). I am not giving legal advice."
Climate: "Federal climate-science pages say recent warming is largely from human greenhouse-gas emissions, mainly burning fossil fuels. Earth's climate has also changed for natural reasons over long periods."
The Pentagon: "The Pentagon is the headquarters of the Department of Defense." The President signed an order last year to rebrand it the Department of War.
Presidential pay: the chatbot said "some presidents have declined to keep that pay," listing George Washington, Herbert Hoover and John F. Kennedy before Trump. At the launch, Trump said he was the first president to forgo a salary.
Then the Answers Changed
According to AP, "as Trump was still speaking at Tuesday's launch event, however, the site began modifying its answers." Asked again about 2020, it replied that it couldn't answer "political questions." It gave the same response to questions about whether Trump had been impeached or convicted of felony charges.
What This Shows About AI
Our view: this isn't really a story about one website. It shows something true of every AI chatbot. Its answers aren't fixed facts pulled from a drawer. They're shaped by the sources it reads and by the rules its operators set, and those rules can change at any moment, without notice.
Commercial chatbots work the same way. Companies decide which topics their AI avoids, how it phrases sensitive answers, and when it refuses. The difference here is that a government AI speaks with official authority.
Why It Matters for Families
This is one of the best real-world media-literacy examples you'll find. Three ideas to share with a teen:
An AI answer is a snapshot. The same question can get a different answer tomorrow. Screenshot and date anything important.
"I can't answer that" is also a choice. A refusal tells you about the rules, not the facts. Ask: who decided this topic was off-limits?
Check the source behind the source. If an AI says it's quoting official pages, go read those pages yourself.
A conversation starter: ask your teen to put the same question to two different chatbots and compare the answers. Then ask: why might they differ?
The Part Worth Remembering
An official AI is only as trustworthy as the rules about what it's allowed to say, and how openly those rules are set and changed.
Source: Associated Press, Meg Kinnard, Chris Rugaber and Will Weissert, "Trump touts America.gov, but its government-sourced AI answers undercut many of his statements," September 29, 2026. https://apnews.com/article/trump-americagov-website-ai-questions-answers-ff86fcb161c1fe89538d6ecb294ea353
Chatbot answers quoted verbatim as reported by AP. The interpretation and family guidance are ours.
Disclosure: drafted with Claude, made by Anthropic.
Want practical AI guidance for parents and educators every week? Subscribe: https://www.aibyage.com/?modal=signup&utm_source=beehiiv&utm_medium=newsletter