Google starts rollout of restrictions on AI as key global election year looms

Google will be rolling out anti-fake news measures across many platforms (Tim Goode/PA)

Google has started the rollout of restrictions on what kind of election-related questions its AI chatbot Gemini will answer as it tries to prevent the spread of fake news during a year when billions of people will vote worldwide.

The technology giant said that users in India will be restricted as to what they can ask Gemini, or at least what types of questions it will provide responses to.

It is part of the company’s efforts to ensure that misinformation and disinformation is limited in a year when according to the Centre for American Progress more than two billion people in 50 countries will head to the polls.

Some of these elections will be contested freely and fairly, while others will not.

The countries where votes are being held this year include the US, Mexico, Russia and probably the UK as well.

But by far the biggest is India, where around 900 million people are registered to vote according to Chatham House.

“With millions of eligible voters in India heading to the polls for the general election in the coming months, Google is committed to supporting the election process by surfacing high-quality information to voters, safeguarding our platforms from abuse and helping people navigate AI-generated content,” Google said in a blog post.

The tech giant laid out a series of non-AI measures it was taking to try to reduce the harm for which its platforms might be used to spread.

These include efforts to provide information directly from the Electoral Commission of India on Google Search and YouTube.

But it will also include restrictions on how Gemini can be used.

“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,” the Google India team said.

“We take our responsibility for providing high-quality information for these types of queries seriously, and are continuously working to improve our protections.”