OpenAI, the pioneer behind ChatGPT, is taking its artificial intelligence model to the next level by introducing AI assistants to help human trainers fine-tune the output of chatbots. The innovative approach, known as reinforcement learning with human feedback (RLHF), has proven successful in making chatbots more reliable and useful, but it has its limitations. Human feedback can be inconsistent, and rating complex outputs, such as software code, can be challenging even for skilled humans. To address these limitations, OpenAI has developed a new model, CriticGPT, which can catch bugs that humans miss and provide better critiques of code. The company plans to extend this approach to areas beyond code in the future. This breakthrough has significant implications for making AI models more accurate and trustworthy, and it could be a crucial step towards creating AI that exceeds human abilities.
AI Trainers Get AI Assistants
OpenAI is integrating this technique into its RLHF chat stack, which could help make its models and tools like ChatGPT more accurate by reducing errors in human training.
1–2 minutes
TOP STORIES
Obama stresses Democrats must prioritize AI safety and economic strategy …
Amodei’s call for caution in AI development highlights the need for safety and alignment …
Sony Music and Warner Music have sued Anthropic, accusing it of copyright infringement in AI training …
Huang believes Nvidia’s position in AI will lead to another year of record growth …
Anthropic’s report reveals a surge in distillation attacks by Chinese AI firms on U.S. models …
AI agents’ autonomy poses significant risks, as demonstrated by a recent breach …
