Google's Gemini AI restricted from answering global election questions

Mountain View: Google has implemented restrictions on its AI chatbot, Gemini, preventing it from responding to questions related to global elections happening in 2024.

This decision aims to prevent potential missteps in deploying the technology amid concerns about misinformation and fake news associated with generative AI advancements.

When users inquire about elections, including the upcoming U.S. presidential match-up between Joe Biden and Donald Trump, Gemini responds, "I'm still learning how to answer this question. In the meantime, try Google Search."

The restrictions were initially announced for the U.S. in December and were set to take effect ahead of the election. The recent update extends these limitations globally in preparation for elections across various countries in 2024.

A Google spokesperson stated, "In preparation for the many elections happening around the world in 2024 and out of an abundance of caution, we are restricting the types of election-related queries for which Gemini will return responses."

Apart from the United States, several large countries, including South Africa and India, are gearing up for national elections. In India, tech firms have been instructed to obtain government approval before publicly releasing AI tools under trial or considered "unreliable" and to label them for the potential to provide incorrect answers.

This move by Google follows scrutiny of its AI products after inaccuracies in historical depictions created by Gemini prompted the company to temporarily halt the chatbot's image-generation feature. CEO Sundar Pichai acknowledged the biases and inaccuracies, stating that the responses were "completely unacceptable" and that corrective measures were underway.

Meta Platforms, the parent company of Facebook, also announced plans to establish a team to combat disinformation and misuse of generative AI leading up to the European Parliament elections scheduled for June.