The article discusses a factual error made by Google’s Gemini AI system during a demo video showcasing its capabilities. Gemini, an AI model trained to engage in open-ended dialogue, incorrectly stated that the Parthenon in Athens, Greece, was constructed in the 16th century. In reality, the Parthenon was built in the 5th century BC, between 447 and 432 BC. This error raised concerns about the potential risks of AI systems providing inaccurate information, even when they appear confident in their responses. The incident highlighted the importance of fact-checking and verifying information from AI models, as they can make mistakes or provide incorrect data. While AI systems like Gemini are impressive in their ability to engage in natural conversations, they are not infallible and can propagate misinformation if not properly vetted. The article serves as a reminder of the need for critical thinking and fact-checking when relying on AI-generated information.
Source: https://www.businessinsider.com/google-gemini-made-factual-error-in-demo-video-2024-5