Google blocks US polling requests in Gemini AI.

Gemini, an AI chatbot from Google, refuses to answer questions about the upcoming US election. The giant of technology he said. On Tuesday, an attempt was made to extend the limits surrounding the earlier elections in India. Reuters report that The ban will increase worldwide.

“As we prepare for the many elections around the world in 2024, and out of an abundance of caution, we're limiting the number of election-related questions that Gemini can answer,” a Google spokesperson told Decrypt—restrictions were first announced in December.

The spokesperson added: “The December blog shows who we work with in the US.

Heading into the heated 2024 election season, AI developers like OpenAI, Anthropic and Google are poised to pounce. Wrong information Using their platforms, but refusing to answer even basic Gemini questions – like the day of the US presidential election – is a new level of moderation.

“Supporting elections is an important part of Google's responsibility to our users and the democratic process,” the company said. “Protecting the integrity of elections means protecting our products and services from attacks.

“Around Google, we have long-standing policies to protect the security of our products and platforms,” ​​the statement continued. Regardless of content type, our policies are consistent and apply to all users.

Asked about the upcoming election, Gemini said, “I'm still learning how to answer that question. In the meantime, try a Google search.

When it comes to Election Day, Google Search gives a straightforward answer: Tuesday, November 5, 2024.

Google did not immediately respond. Decrypt's Request for feedback.

Asking Gemini's rival OpenAI's ChatGPT the same question. GPT-4 “The 2024 United States presidential election is scheduled for Tuesday, November 5, 2024.

OpenAI is not accepted. Decrypt's Ask for a comment, instead pointing to the January blog Post About the company's approach to elections around the world.

“We expect and aim for people to use our devices safely and responsibly, and polls are no different,” OpenAI said. “We work to anticipate and prevent related attacks, such as deceptive “deep fakes,” overflowing influencers, or chatbots impersonating candidates.”

For its part, Anthroponic has officially announced Cloud AI. Unlimited For political candidates. Again, the cloud highlights not only the election date, but other election-related information.

“We don't allow candidates to use the cloud to build chatbots that can impersonate them, and we don't allow anyone to use the cloud for targeted political campaigns.” Anthropic he said. last month. We have trained and deployed automated systems to detect and prevent misuse of information such as misinformation.

Anthroponic says that violating the company's voting restrictions may result in the suspension of a user's account.

“Because generative AI systems are relatively new, we are taking a cautious approach to how our systems are used in politics,” Anthropic said.

Edited by Ryan Ozawa.