Google and Alphabet CEO Sundar Pichai delivers a speech on artificial intelligence at the Bruegel Think Tank Conference in Brussels, Belgium, on January 20, 2020.
Yves Herman | Reuters
Google announced it will limit the types of election-related queries users can ask its Gemini chatbot, adding that it has rolled out the changes in India, where voters will head to the polls this spring.
“Out of an abundance of caution on such an important topic, we have begun limiting the types of election-related queries to which Gemini will return responses,” Google wrote in a statement. Blog article Tuesday. “We take our responsibility for providing high-quality information to inquiries like this seriously and are continually working to improve our safeguards.”
A Google spokesperson told CNBC that the changes are in line with how the company plans to hold elections, and that Gemini restrictions were introduced “in preparation for the many elections taking place around the world in 2024 and out of an abundance of caution.”
The announcement comes after Google withdrew its artificial intelligence image generation tool last month amid a series of controversies, including historical inaccuracies and controversial responses. As part of a major rebranding, the company launched its image generator in early February via Gemini, Google’s main AI model suite.
“We have taken the feature offline while we fix the issue,” Google DeepMind CEO Demis Hassabis told a panel at Mobile World Congress in Barcelona last month. “We have taken the feature offline while we fix the issue.” Hopefully it will be back online soon in the next few weeks.” He added that the product “wasn’t working the way we expected.”
The news comes as technology platforms prepare for a crucial election year affecting more than 4 billion people in more than 40 countries around the world. The rise of artificial intelligence-generated content has raised serious concerns about election-related misinformation, with the amount of deep fake content generated increasing 900% year-on-year, according to data from machine learning company Clarity.
Election-related misinformation has been a major problem dating back to the 2016 presidential campaign, when Russian actors seeking deployment A cheap and easy way to spread inaccurate content on social platforms. Lawmakers are now more concerned about the rapid rise of artificial intelligence.
“There are legitimate concerns about how artificial intelligence can mislead voters in campaigns,” California Democratic state Sen. Josh Becker told CNBC last month.
Detection and watermarking technologies for identifying deep fakes are not developing fast enough to keep up. Even if the platforms behind AI-generated images and videos agree to bake in invisible watermarks and certain types of metadata, there are ways to bypass these protections. Sometimes, screenshots can even fool detectors.
In recent months, Google has emphasized its commitment to pursue and invest heavily in artificial intelligence assistants, or agents. The term often describes tools ranging from chatbots to coding assistants and other productivity tools.
Alphabet Chief Executive Sundar Pichai emphasized artificial intelligence agents as a priority during the company’s earnings call on Jan. 30. Pichai said he ultimately hopes to provide an artificial intelligence agent that can complete an increasing number of tasks for users, including completing tasks in Google searches – although he said there is “a lot of execution work ahead.” Likewise, CEOs of tech giants from Microsoft to Amazon are doubling down on turning artificial intelligence agents into productivity tools.
Sissie Hsiao, Google vice president and general manager of Google Assistant and Bard, told reporters on a conference call in February that Google’s Gemini rebranding, application launch and feature expansion are the first steps in “creating a true artificial intelligence assistant.”