OpenAI's research reveals that ChatGPT can produce harmful gender or racial stereotypes based on user names, with a bias rate of about one in 1000 responses. Despite these rates appearing low, the vast user base of 200 million weekly interactions means that even small percentages can lead to significant bias. The study highlights the need for improved evaluation methods to address these biases in AI models.
The research distinguishes between 'first-person fairness' and 'third-person fairness,' focusing on how user names influence ChatGPT's responses. Examples show that responses can vary significantly based on perceived gender, indicating a reliance on historical stereotypes. OpenAI aims to expand its analysis to include various user attributes, enhancing the understanding of bias in AI interactions.
• ChatGPT shows bias based on user names in responses.
• OpenAI aims to improve AI model evaluations to reduce bias.
First-person fairness refers to how AI responses are influenced by the user's name.
Third-person fairness involves bias in AI when screening resumes or loan applications.
RLHF is a training method where human testers guide AI to improve response quality.
OpenAI develops AI models like ChatGPT and is focused on reducing bias in AI interactions.
Google is mentioned as having similar bias rates in its AI models, like Gemini.
Cointelegraph.com on MSN.com 12month
Isomorphic Labs, the AI drug discovery platform that was spun out of Google's DeepMind in 2021, has raised external capital for the first time. The $600
How to level up your teaching with AI. Discover how to use clones and GPTs in your classroom—personalized AI teaching is the future.
Trump's Third Term? AI already knows how this can be done. A study shows how OpenAI, Grok, DeepSeek & Google outline ways to dismantle U.S. democracy.
Sam Altman today revealed that OpenAI will release an open weight artificial intelligence model in the coming months. "We are excited to release a powerful new open-weight language model with reasoning in the coming months," Altman wrote on X.