MIT Study Reveals AI Chatbots Show Bias Against Vulnerable Users
MIT research finds GPT-4, Claude 3 Opus, and Llama 3 provide less accurate responses to non-native English speakers and less educated users.
MIT research finds GPT-4, Claude 3 Opus, and Llama 3 provide less accurate responses to non-native English speakers and less educated users.
DeepSeek and Alibaba AI models spreading globally are programmed to present positive views of China while avoiding negative topics, researchers discover.
Tests reveal that OpenAI's latest ChatGPT model is citing Elon Musk's AI-generated encyclopedia, Grokipedia, as a source, leading to concerns about the spread of misinformation and biased narratives.