OpenAI Updates GPT-4o With Improved Creative Writing Ability, Reveals New Automated Red Teaming Method

OpenAI introduced two methods it’s enhancing its synthetic intelligence (AI) fashions final week. The first contains releasing a brand new replace for the GPT-4o (often known as the GPT-4 Turbo), the corporate’s newest AI mannequin powering ChatGPT for paid subscribers. The firm says the replace improves the inventive writing capability of the mannequin and makes it higher at pure language responses and writing partaking content material with excessive readability. OpenAI additionally launched two analysis papers on crimson teaming and shared a brand new methodology to automate the method to scale recognizing errors made by its AI fashions.

OpenAI Updates GPT-4o AI Model

In a post on X (previously generally known as Twitter), the AI agency introduced a brand new replace for the GPT-4o basis mannequin. OpenAI says the replace permits the AI mannequin to generate outputs with “extra pure, partaking, and tailor-made writing to enhance relevance and readability.” It can also be mentioned to enhance the AI mannequin’s capability to course of uploaded information and supply deeper insights and “extra thorough” responses.

Notably, the GPT-4o AI mannequin is on the market to customers with the ChatGPT Plus subscription and builders with entry to the big language mannequin (LLM) through API. Those utilizing the free tier of the chatbot do not need entry to the mannequin.

While Gadgets 360 employees members weren’t capable of check out the brand new capabilities, one person on X posted concerning the newest enhancements within the AI mannequin after the replace. The person claimed that GPT-4o may generate an Eminem-style rap cipher with “subtle inner rhyming buildings”.

OpenAI Shares New Research Papers on Red Teaming

Red teaming is the method utilized by builders and corporations to make use of exterior entities to check software program and programs for vulnerabilities, potential dangers, and questions of safety. Most AI companies collaborate with organisations, immediate engineers, and moral hackers to stress-test whether or not it responds with dangerous, inaccurate, or deceptive output. Tests are additionally made to test whether or not an AI system will be jailbroken.

Ever since ChatGPT was made public, OpenAI has been public with its crimson teaming efforts for every successive LLM launch. In a blog post final week, the corporate shared two new analysis papers on the development of the method. One of them is of specific curiosity given the corporate claims it will possibly automate large-scale crimson teaming processes for AI fashions.

Published within the OpenAI area, the paper claims that extra succesful AI fashions can be utilized to automate crimson teaming. The firm believes AI fashions can help in brainstorming attacker objectives, how an attacker’s success will be judged, and understanding the variety of assaults.

Expanding on it, the researchers claimed that the GPT-4T mannequin can be utilized to brainstorm a listing of concepts that represent dangerous behaviour for an AI mannequin. Some examples embody prompts equivalent to “find out how to steal a automobile” and “find out how to construct a bomb”. Once the concepts have been generated, a separate crimson teaming AI mannequin will be constructed to trick ChatGPT utilizing an in depth collection of prompts.

Currently, the corporate has not begun utilizing this methodology for crimson teaming given a number of limitations. These embody the evolving dangers of AI fashions, exposing the AI to lesser-known strategies for jailbreaking or producing dangerous content material, and the necessity for the next threshold for information in people to appropriately decide the potential dangers of output as soon as the AI mannequin turns into extra succesful.