Google has declared its intention to limit certain election-related inquiries that users can make to its Gemini chatbot. The restrictions have been implemented in the U.S. and India, ahead of upcoming elections in these regions.
“In recognition of the significance of this topic, we’re starting to impose restrictions on the types of election-related queries that Gemini will respond to,” Google stated in a blog post on Tuesday.
“We take our responsibility to provide accurate information for such queries seriously and are continuously enhancing our safeguards.”
A Google spokesperson informed CNBC that these changes align with the company’s pre-planned strategy for election periods.
“As we announced last December, in anticipation of numerous elections worldwide in 2024 and as a precautionary measure, we are limiting the types of election-related queries for which Gemini will provide responses,” the spokesperson explained.
This announcement follows Google’s decision to withdraw its AI-powered image generation tool last month due to various controversies, including historical inaccuracies and contentious responses.
The tool was introduced in February via Gemini, Google’s primary suite of AI models, as part of a significant rebranding effort.
“We have temporarily disabled the feature while we address these issues,” remarked Demis Hassabis, CEO of Google’s DeepMind, during a panel discussion at the Mobile World Congress conference in Barcelona.
“We hope to restore functionality in the next couple of weeks. The product did not perform as expected.”
This development coincides with tech platforms gearing up for a significant year of elections worldwide, affecting over four billion people in more than 40 countries.
The proliferation of AI-generated content has raised serious concerns about election-related misinformation, with the number of deepfakes generated increasing by 900% year over year, according to data from machine learning firm Clarity.
Election-related misinformation has been a persistent issue since the 2016 presidential campaign when Russian actors utilized inexpensive methods to disseminate inaccurate content across social media platforms.
Lawmakers are now even more concerned about the rapid advancement of AI technology.
“There is a genuine reason for concern regarding how AI could be manipulated to deceive voters during campaigns,” remarked Josh Becker, a Democratic state senator in California, during an interview with CNBC last month.
The technologies used to detect and watermark deepfakes have not progressed rapidly enough to keep pace.
Even if platforms incorporating AI-generated images and videos agree to embed invisible watermarks and specific metadata, there are ways to circumvent these protective measures. In some cases, a simple screenshot can deceive detection systems.
Google has reiterated its commitment to advancing and investing in AI assistants or agents in recent months. These tools encompass a wide range, from chatbots to coding assistants and other productivity aids.
Alphabet CEO Sundar Pichai underscored the importance of AI agents during the company’s earnings call on January 30.
Pichai expressed his desire to offer an AI agent capable of handling an increasing array of tasks for users, including within Google Search, though he acknowledged that significant execution challenges lie ahead.
Similarly, CEOs of tech giants such as Microsoft and Amazon have reaffirmed their dedication to developing AI agents as productivity tools.
Sissie Hsiao, a vice president at Google and general manager for Google Assistant and Bard, characterized Google’s Gemini rebrand, app launches, and feature expansions as the initial steps toward “creating a true AI assistant” during a call with reporters in February.
