Google AI Mode declines to name 2020 election winner

The AI's non-committal response

Google's AI Mode, which is the AI-generated summary feature built into Google Search, recently drew scrutiny for its handling of the 2020 presidential election. When asked who won the election, the tool did not provide a direct answer naming the winner. Instead, it provided a link to an official document while avoiding a definitive statement.

The response included a link to a PDF titled "OFFICIAL 2020 PRESIDENTIAL GENERAL ELECTION RESULTS." This document lists the certified totals showing Joe Biden secured 306 Electoral College votes while Donald Trump received 232. Despite these clear figures, the AI summary itself remained silent on the final outcome.

This behavior was notable because the tool behaved differently for other election cycles. Users documented that the AI Mode correctly named Donald Trump as the winner for both the 2016 and 2024 presidential elections. The inconsistency regarding 2020 led to immediate public interest and scrutiny of the underlying model.

Google's official stance

Google addressed the incident by stating that its standalone chatbot, Gemini, provides more complete answers to the same question. The company noted that the output from AI Mode was less useful than the version provided by Gemini. However, Google has not provided a specific technical explanation for why the two products produced different results.

The company has also not confirmed any outside involvement in shaping the specific response. While some users have speculated about external influences, Google has not elaborated on the cause of the discrepancy. The AI Mode response included a standard disclaimer warning users that artificial intelligence can make mistakes and should be double-checked.

A Google spokesman declined to answer questions regarding whether the Trump administration had any role in shaping the chatbot's response. This lack of detail comes as Republican members of the House of Representatives have urged Alphabet to reveal any potential government influences on the development of Gemini. The company remains silent on these specific allegations of external pressure.

A shifting landscape for news

The timing of this incident is significant as it occurred just days before the November 3 midterm elections. This period is characterized by ongoing public debate regarding the integrity of American elections. Donald Trump has continued to cast doubt on the 2020 results, despite a lack of credible evidence of interference from public records and officials.

The incident highlights a major shift in how the public consumes information. A study commissioned by the firm Orchestra shows that the percentage of American adults using chatbots is expected to climb from roughly one-third in 2024 to half of the population by 2026. More people are increasingly relying on AI-generated summaries to find facts about current events and historical data.

Because AI tools are becoming primary gateways for information, the accuracy and neutrality of these summaries are under intense pressure. When a tool provides a link to a fact but refuses to state the fact itself, it creates a complex user experience. This dynamic is especially sensitive during election cycles where information accuracy is a high priority for the public.

Past disputes and future updates

This incident follows a series of disagreements regarding how Google's AI products manage political topics. Google previously confirmed it would restrict its Gemini chatbot from answering election-related queries globally. This update was already live in the United States at the time of the AI Mode incident, showing a broader effort to manage sensitive content.

Google CEO Sundar Pichai has previously acknowledged that some of the chatbot's responses have offended users and shown bias. He described such instances as "completely unacceptable" and has pushed for improvements in the technology. These statements reflect the ongoing challenge of balancing helpfulness with the need to avoid generating biased or incorrect political content.

While Newsweek reported that Google AI gave equivalent answers for 2020 and 2024 as of October 19, the initial discrepancy remains a point of concern. The company continues to refine its AI Mode to ensure it provides clear and helpful summaries. For now, the discrepancy between the AI Mode and the Gemini chatbot highlights the complexities of deploying large-scale AI in a political context.