Meta, OpenAI, Anthropic and Cohere A.I. models all make stuff up — here's which is worst
Publishing timestamp: 2023-08-18 08:35:28
Summary
The article discusses a report by researchers at Arthur AI who tested top AI models from Meta, OpenAI, Cohere, and Anthropic. The researchers found that some models "hallucinate" or make up facts more than others. GPT-4 performed the best overall and hallucinated less than its prior version, GPT-3.5. Cohere's AI model hallucinated the most. The report highlights the importance of testing AI models on specific workloads to understand their performance in real-world applications.
Sentiment: NEUTRAL
Keywords: breaking news: technology, business news, arthur j. gallagher & co., technology, generative ai, microsoft corp, science, artificial intelligence, meta platforms inc,
Source: https://www.cnbc.com/2023/08/17/which-ai-is-most-reliable-meta-openai-anthropic-or-cohere.html