A new study from Forum AI found that ChatGPT, Gemini, Claude, and Grok provide inaccurate answers when asked about elections and geopolitical topics. The findings raise concerns about AI reliability on critical information.
Four major chatbots—OpenAI's ChatGPT, Google's Gemini, Anthropic's Claude, and xAI's Grok—struggle to deliver fair and accurate responses on elections and news-related questions, according to research by Forum AI.
The startup's survey revealed a pattern of unreliable outputs across the industry's leading AI models when handling sensitive topics like geopolitics and electoral information. This comes as these chatbots gain widespread adoption for news consumption and information retrieval.
Forum AI CEO Campbell Brown attributed the problem to a systemic issue within the AI industry: companies are essentially "grading their own homework" when it comes to evaluating their models' performance. She argued this self-assessment approach lacks the necessary external oversight to catch and correct accuracy problems before deployment.
The findings underscore growing tensions between AI capabilities and real-world reliability. As chatbots become increasingly integrated into how people access news and political information, their propensity for inaccurate answers poses potential risks to informed decision-making, particularly during election cycles.
The study does not specify the exact nature of the inaccuracies—whether chatbots hallucinate entirely false information, present biased perspectives, or conflate multiple events. However, the broad finding suggests all four major models share common accuracy gaps on these high-stakes topics.
Industry observers point to the challenges of training large language models on complex, contextual information with multiple valid interpretations. News and election coverage inherently involves nuance, and chatbots have historically struggled with topics requiring real-time updates or detailed factual precision.
Brown's call for independent evaluation reflects a broader push for third-party auditing of AI systems before public release. As regulators worldwide consider AI governance frameworks, transparency in model performance—particularly on societal topics—has become a key regulatory concern.
The research highlights a critical gap between public perception of chatbot reliability and their actual performance on consequential subjects.
Uber's weekly AI agent requests have grown nearly tenfold since February, yet the company has held spending flat since April after exhausting its entire 2026 AI budget in Q1.
A recent paper shows artificial intelligence often diagnoses and treats patients better than human physicians. The findings are prompting difficult conversations within the medical community about the profession's evolving role.
The Relay Q, launching next year, represents the latest push to establish voice as the primary interface for human-computer interaction, challenging the keyboard's decades-long dominance.
An Anthropic researcher demonstrated automated systems that can identify and correct misaligned behaviors without compromising overall performance. The systems improved on all 10 tested benchmarks measuring specific problematic outputs.