- tech360.tv
Google's Gemini AI Chatbot Restricted from Global Election Queries: Tech Giants Address Misinformation Concerns
Google restricts AI chatbot Gemini from answering queries on global elections. Concerns of misinformation and fake news prompt governments to regulate generative AI. Gemini responds to election-related questions with a suggestion to use Google Search.
In an effort to avoid potential missteps in the deployment of technology, Google has announced that its AI chatbot, Gemini, will be restricted from answering questions about global elections. This decision comes amidst concerns of misinformation and fake news surrounding advancements in generative AI. Governments around the world have been prompted to regulate the technology to address these concerns.
Gemini's response to questions about upcoming elections, such as the U.S. presidential match-up between Joe Biden and Donald Trump in November, is: "I'm still learning how to answer this question. In the meantime, try Google Search." This restriction was initially announced by Google in December of last year, specifically for the U.S. elections. The company spokesperson stated, "In preparation for the many elections happening around the world in 2024 and out of an abundance of caution, we are restricting the types of election-related queries for which Gemini will return responses."
Apart from the U.S., several major countries, including South Africa and India, are scheduled to hold national elections. India has taken a proactive approach by requiring tech firms to seek government approval before releasing AI tools that are deemed "unreliable" or still under trial. Additionally, these tools must be labeled for their potential to provide incorrect answers.
Google's AI products have faced scrutiny recently due to inaccuracies in historical depictions created by Gemini. This led to the suspension of the chatbot's image-generation feature. Sundar Pichai, CEO of Google, acknowledged the issues and referred to the chatbot's responses as "biased" and "completely unacceptable." The company is actively working to address these concerns.
In response to the growing concerns of disinformation and abuse of generative AI, Meta Platforms, the parent company of Facebook, has announced the establishment of a team dedicated to tackling these issues ahead of the European Parliament elections in June. Researchers have warned that AI-powered image creation tools from companies like OpenAI and Microsoft can be used to produce misleading content related to elections or voting.
The Center for Countering Digital Hate (CCDH), a nonprofit organization monitoring online hate speech, conducted tests using AI tools from OpenAI, Microsoft, and Stability AI. The results showed that these tools were susceptible to generating misleading images when prompted with requests related to election fraud. CCDH emphasised the potential for AI-generated images to spread false claims and undermine the integrity of elections.
While some tools successfully blocked prompts requesting images of candidates, others performed poorly. Midjourney, in particular, generated misleading images in 65 percent of the tests conducted by researchers. CCDH also found evidence of people using Midjourney to create misleading political content.
In response to these findings, the tech companies involved have taken steps to address the issue. OpenAI, Microsoft, and Stability AI were among the 20 companies that signed an agreement to work together in preventing deceptive AI content from interfering with elections worldwide. Midjourney, however, was not part of the initial group of signatories.
As the global elections approach, it is crucial for tech companies to take responsibility and ensure the responsible use of AI tools. The potential impact of AI-generated content on the integrity of elections cannot be underestimated. Governments, organizations, and individuals must work together to combat misinformation and safeguard the democratic process.
Google restricts AI chatbot Gemini from answering queries on global elections
Concerns of misinformation and fake news prompt governments to regulate generative AI
Gemini responds to election-related questions with a suggestion to use Google Search
Source: CGTN