OpenAI recently introduced CriticGPT, a new AI tool based on GPT 4 to find errors in code produced by ChatGPT. This latest innovation is part of an effort to improve AI systems to give accurate answers. According to the company, CriticGPT improves code review results by 60 percent compared to those that did not do so. OpenAI will soon integrate CriticGPT into its Reinforcement Learning from Human Feedback (RLHF) labeling pipeline. This move will provide AI trainers with better tools to evaluate AI outputs, no matter how complex they are.

The company shared its research paper which states that CriticGPT is intended to assist human AI reviewers in the task of reviewing code generated by ChatGPT. CriticGPT was built using GPT-4 LLM, and showed encouraging abilities in analyzing code and identifying errors, helping its human colleagues recognize AI “confusions” that they cannot see themselves.

OpenAI explained that for the training process, human developers were asked to edit the code written by ChatGPT, introduce a series of errors, and provide sample feedback. This taught CriticGPT to identify the most common and some less common coding errors.

After the training was completed, OpenAI said the results were impressive. However, one of the challenges that CriticGPT has to overcome is that some AI confusions are the result of errors spread across many different code strings, making it very difficult for CriticGPT to identify the source of the problem.

This tool is one of many new tools being developed to improve large language models and make their generative AI more capable.

The new technology can also help the upcoming AI models developed by OpenAI. The company has recently shared that they are working on their new and more powerful AI models. OpenAI CTO Meera Murati has recently shed light on the upcoming AI model and its intelligence.

Murati says that GPT-3 had the intelligence of a young child, GPT-4 had the intelligence of a high-school kid, and the next generation of the model will have the intelligence of a person with a PhD for specific tasks. He also revealed that the next generation of GPT can be expected in “a year and a half”. He also said that when you talk to the next generation chatbot, in many things you may feel that it is smarter than you.

To make its new generative model even smarter, OpenAI has partnered with Time Magazine. The artificial intelligence research company has teamed up with Time Magazine to enhance and train ChatGPT. On Thursday, the companies announced a multi-year content deal that will allow OpenAI to access current and acquired articles from over 100 years of Time’s work. The new partnership furthers Time’s commitment to expanding global access to accurate and reliable information.