Washington, D.C. – June 3, 2025 – A growing number of reports from media and fact-checking agencies are raising serious concerns over the reliability of artificial intelligence chatbots such as OpenAI’s ChatGPT, Google’s Gemini, and xAI’s Grok in verifying real-time information. As users increasingly turn to AI for fact-checking during fast-moving global events, researchers warn that the tools are not only frequently inaccurate but can also generate and reinforce false narratives.
During the recent four-day conflict between India and Pakistan, social media users queried Grok, the AI assistant integrated into Elon Musk’s platform X, for verification of viral videos. The chatbot incorrectly identified old video footage from Sudan as a missile strike on Pakistan and claimed a fire in Nepal was footage of a Pakistani military response. Similar incidents have occurred with other AI systems, which have offered fabricated details or confirmed the authenticity of AI-generated content.
According to a study by the disinformation watchdog NewsGuard, ten leading AI chatbots, including ChatGPT and Gemini, routinely repeated falsehoods and, in many cases, failed to flag content as misleading. Meanwhile, a Columbia University Tow Centre study concluded that chatbots often fabricate answers when they cannot confidently provide correct information, rather than choosing to remain silent.
Experts point to the reduction in human-led fact-checking teams at major tech companies like Meta and X as a contributing factor to the growing reliance on AI for information verification. In the absence of dedicated professionals, platforms are shifting toward community-driven models like “Community Notes,” which researchers argue are insufficient for handling complex or politically sensitive misinformation.
The controversy has been further inflamed by reports that AI responses can be manipulated or biased based on how systems are programmed. Grok, for example, recently inserted references to the “white genocide” conspiracy theory in unrelated user queries, which xAI later attributed to an unauthorized prompt modification. When asked who could have caused the prompt change, Grok named Elon Musk as the most likely individual—a response that has sparked alarm over transparency and accountability in AI systems.
Concerns about political influence on chatbot output continue to grow, especially in light of recent incidents where AI tools reinforced Russian and Chinese disinformation narratives, as well as fabrications during the Australian election. Global fact-checking organizations, including AFP and the International Fact-Checking Network, have called for renewed investment in human moderation and greater transparency in AI development to ensure public trust is not eroded further.
As AI continues to shape the future of information dissemination, experts warn that caution and human oversight are more critical than ever. Inaccurate or biased AI responses, especially during high-stakes events, not only risk misleading the public but can also destabilize democratic discourse.