Google says it will restrict Gemini's ability to answer election-related questions 'out of an abundance of caution'

• Google will restrict its AI chatbot Gemini from answering election-related questions in 2024.
• This comes amid the biggest global election year ever, with at least 64 countries holding elections.
• The decision follows global concerns over the potential misuse of AI in elections.

Google will restrict its AI chatbot Gemini from answering questions about elections as a precautionary measure and "out of an abundance of caution," a company spokesperson said.

This year will be the biggest global election year in history, with over two billion voters set to go to the polls, according to The World Economic Forum. At least 64 countries will hold elections.

"In preparation for the many elections happening around the world in 2024 and out of an abundance of caution, we're restricting the types of election-related queries for which Gemini will return responses," a Google spokesperson told Business Insider by email.

Gemini already refuses to answer questions about the upcoming US presidential election, where President Joe Biden and former President Donald Trump are set to go head-to-head.

It responds to election-related questions with a generic message: "I'm still learning how to answer this question. In the meantime, try Google Search."

The news of how Gemini will handle election-related inquiries comes just a couple of weeks after concerns were raised over its image-generating capability.

In late February, Google pulled Gemini's ability to generate images of people following complaints on social media that it produced images of people of color in historically inaccurate contexts.

In India, the world's largest democracy, which has an election in the spring, Google confirmed in a blog post that it had already started rolling out restrictions on the types of questions Gemini will answer.

"We take our responsibility for providing high-quality information for these types of queries seriously, and are continuously working to improve our protections," the blog post said.

On Tuesday, the Indian government also asked tech companies to obtain its approval before publicly launching "unreliable" or "under-tested" generative AI models or tools, Al Jazeera reported.

Representatives from OpenAI, which developed ChatGPT, met with officials from the Election Commission of India last month to discuss how the AI tool could be used safely in the election, The Hindu reported last week.

Former FBI director Christopher Wray warned security professionals last month that advancements in AI could make it easier for "less sophisticated foreign adversaries to engage in malign influence," which could make foreign influence efforts harder to detect.