#openai

According to insights from the Wall Street Journal, Meta's ongoing project is a new AI system. They aim to topple established language models, such as GPT-4 and even their own Llama 2. This fresh initiative will center on large-scale text processing and its subsequent scrutiny. The timeline for initiating the model's training is set for 2024.

Social media giants, like Facebook, rely on a vast global moderation teams to sift out damaging content, from child pornography materials to severe violence. The implementation of GPT-4 might be the key to revolutionizing this workflow. "The process [of content moderation] is inherently slow and can lead to mental stress on human moderators. With this system, the process of developing and customizing content policies is trimmed down from months to hours," assert OpenAI representatives. Access to the OpenAI API now enables individuals to forge their distinct AI-driven moderation frameworks.

OpenAI and Microsoft are under fire in a class-action lawsuit over alleged web scraping to gather private data for training AI models like ChatGPT. The lawsuit, filed on Sept. 5 in San Francisco, accuses the companies of using stolen private info, including personal data from countless internet users, without consent. The plaintiffs argue that without this data theft, products like ChatGPT and DALL-E wouldn't be the billion-dollar businesses they are today. They seek damages and a halt to the alleged illegal data scraping.

According to OpenAI’s blog post, web pages crawled with GPTBot could potentially be utilized to enhance accuracy and expand the capabilities of upcoming iterations of their AI models. A web crawler (web spider), is a type of bot that indexes website content across the internet. Popular search engines like Google and Bing employ web crawlers to enable websites to appear in search results. OpenAI clarified that their web crawler will gather publicly available data from the world wide web while excluding sources with paywalled content, personally identifiable information, or text that violates their policies.

Tech leaders Google, Microsoft, OpenAI, and AI startup Anthropic are forming a new industry watchdog group, the Frontier Model Forum, to supervise the "safe and responsible" development of frontier AI models. The alliance aims to push research on AI safety, pinpoint responsible development practices, collaborate with governments and civil leaders, and back the development of applications. This move follows a commitment made by these AI companies to the White House for the safe and transparent development of AI.