Post-Cutoff

Public reportMedium severity

Gemini app called Donald Trump the “former president” six weeks into his second term

Gemini app3 March 2025, public report

What was trueDonald Trump was the sitting president, six weeks into his second term, and J. D. Vance was vice president. Source: techcrunch.com

What happened

TechCrunch (Maxwell Zeff, 4 Mar 2025) tested Gemini on political questions. Other chatbots answered them.

What was reported (verbatim from TechCrunch)

As of Monday morning, Gemini demurred when asked to identify the sitting U.S. president and vice president, according to TechCrunch’s testing.

In one instance during TechCrunch’s tests, Gemini referred to Donald J. Trump as the “former president” and then declined to answer a clarifying follow-up question.

Google spokesperson, quoted verbatim:

“Large language models can sometimes respond with out-of-date information, or be confused by someone who is both a former and current office holder,” … “We’re fixing this.”

Why this is striking

  • The vendor itself confirms out-of-date information as the cause.
  • The wrong answer came from the model’s pre-2025 prior (Trump = former president). It wasn’t a guess about something unknown.

Correction

Late Monday, after TechCrunch alerted Google of Gemini’s erroneous responses, Gemini started to correctly answer that Donald Trump and J. D. Vance were the sitting president and vice president … However, the chatbot wasn’t consistent, and it still occasionally refused to answer the questions.

Lesson

Facts about current officeholders are exactly what goes out of date after a cutoff. Vendors end up handling them in system prompts: the published Claude Opus 5.5 system prompt says that for “current news or events (e.g. current officeholders)” Claude gives its most recent pre-cutoff information, “notes it may be outdated, and points to web search”.

Sources

  • TechCrunch, “Google still limits how Gemini answers political questions” (2025-03-04): techcrunch.com

Related cases

  1. Our testCritical severity

    All four Gemini and Claude runs called the briefing speculative fiction

  2. Our testHigh severity

    Gemini 3.8 Flash labelled real, released models and real events as “fictional”

  3. Our testMedium severity

    Gemini 3.8 Flash called real, announced Apple products “conceptual product reveals”

Spotted an error in this case file? Write to contact@postcutoff.com. All cases