OpenAI announced it started working on a new product that will bring humanity closer to Artificial General Intelligence.
OpenAI, the tech start-up behind ChatGPT and DALL-E, said it began working on a new product that will bring humanity closer to Artificial General Intelligence (AGI). The firm also announced the creation of a new safety team to oversee OpenAI’s operations.
The previous safety team was dissolved in May, the last of a month-long shake-up in the executive positions of the start-up.
In October last year, CEO Sam Altman was ousted by OpenAI’s board and then reinstated after a few days. The reason for this sudden move remains unclear to this day, with some speculating the board was unsatisfied with the profit-driven approach implemented by Altman.
Though technically non-profit, OpenAI is partially owned by Microsoft. The tech giant acquired part of the firm in a $10 billion deal last year and has since focused heavily on AI development.
After Altman’s swift return, both co-founder Ilya Suskever and key researcher Jan Leike resigned from the board. Leike denounced that “safety culture and processes have taken a backseat to shiny products” in OpenAI’s operations.
Sam Altman said in response that he was sad to see Leike leave the company.
The new safety team was announced in a blog post. It will consist of Sam Altman, Bret Taylor, Adam D’Angelo, and Nicole Seligman, all members of OpenAI’s board. In addition, the oversight team will include four technical experts already working at the company.
The team will review OpenAI’s processes and report their findings to the board within 90 days. Their recommendations will also be shared in a future blog post.
OpenAI’s new product
Also in a blog post on Tuesday, OpenAI announced it began training on a new generative AI model. This "next frontier model" will bring "the next level of capabilities" to the technology, the post added.
The company has been actively striving toward AGI since its inception. Artificial General Intelligence is a buzzword that means an AI is intelligent or more than humans. Essentially, an artificial intelligence capable of doing several tasks at once and combining them into thought.
ChatGPT, OpenAI’s flagship product, is not even close to the general idea of AGI. ChatGPT is a generative chatbot based on predictive algorithms: based on an immense amount of data, the software predicts which word is more likely to come next in a complex sentence. Although a powerful and impressive technology, ChatGPT does not understand language or reason, and cannot do much more other than create sentences.
However, many are optimistic that OpenAI can actually lead humanity to the creation of AGI. And, according to OpenAI, the new model will be a step closer to that.
