On March 12, 2024, Google announced that it would restrict its AI chatbot Gemini from answering many election-related questions, ahead of the elections held in many countries that year, including the presidential election in the United States. The move was meant to reduce the risk of the chatbot spreading errors or misinformation about elections.
Updated October 2026. This is a news report from March 2024. We added what happened next with the restrictions and with Gemini’s image generation of people.
Generative AI tools, such as image and video generators, have made it easier to spread misinformation and fake news, and governments and election authorities have been pressing technology companies to act.
In December 2023, Google introduced these limitations in the United States, stating that they would be implemented during the presidential elections in 2024. Reportedly, Google’s AI chatbot Gemini responded to a question about the presidential race between Joe Biden and Donald Trump, saying, “I’m still learning how to answer this question. In the meantime, try Google Search.”
Alongside the United States, 2024 also saw elections in other major countries, including the UK, South Africa, Russia, and India, the world’s biggest democracy.
Google India posted a blog on Tuesday that stated,
Out of an abundance of caution on such an important topic, we have begun to roll out restrictions on the types of election-related queries for which Gemini will return responses. We take our responsibility for providing high-quality information for these types of queries seriously, and are continuously working to improve our protections.
The announcement comes just days after Google admitted messing up the image generation feature in its artificial intelligence model. After users reported that Gemini was responding with inaccurate and offensive images, the tech giant stated in a blog post, “This wasn’t what we intended.”
This is not the first restriction on Gemini. Following the inaccurate results, Google stated,
So we turned the image generation of people off and will work to improve it significantly before turning it back on. This process will include extensive testing.
What happened next
The election limits outlasted many of the votes they were meant to cover. In June 2024, after India’s election had ended, TechCrunch reported that Meta had lifted a similar block in Meta AI in India, while Google was still restricting election-related queries in Gemini worldwide and pointing users to Google Search instead. Google did not say if or when it would lift the restrictions.
Image generation of people returned later that year. On August 28, 2024, Google announced that its Imagen 3 model was coming to Gemini and that it was starting to roll out image generation of people again, first to Gemini Advanced, Business and Enterprise users in English. Google said it still does not support generating photorealistic images of identifiable individuals, images of minors, or excessively gory, violent or sexual scenes.
Google’s policies for election questions can change from one election to the next, so if you use Gemini for election information, check official election authorities and trusted news sources as well. For more on Gemini, see The AI War – ChatGPT Vs. Gemini AI: Which One Makes the Wiser Choice?.
