OpenAI says it’s training a successor to GPT-4

OpenAI is coaching a brand new prime AI mannequin to succeed its present GPT-4.

First reported by The New York Instances and introduced in a weblog submit, the corporate is engaged on a successor to the bogus intelligence mannequin that fuels its well-known chatbot, ChatGPT.

“OpenAI has lately begun coaching its subsequent frontier mannequin and we anticipate the ensuing techniques to deliver us to the following stage of capabilities on our path to AGI,” reads the submit, printed Tuesday.

OpenAI’s GPT-4 was launched in March 2023, a robust replace to GPT-3 which powered its profitable ChatGPT chatbot launched in November 2022.

Mashable reached out to OpenAI for additional info on the brand new AI mannequin, and we had been pointed to the weblog.

In the identical submit, OpenAI additionally introduced it has shaped a Security and Safety Committee that “might be liable for making suggestions to the total Board on vital security and safety choices for OpenAI tasks and operations.”

The committee might be comprised of administrators Bret Taylor, Nicole Seligman, Adam D’Angelo, and OpenAI CEO Sam Altman.

“Whereas we’re proud to construct and launch fashions which can be industry-leading on each capabilities and security, we welcome a strong debate at this essential second,” OpenAI’s submit reads.

The information comes weeks after OpenAI unveiled GPT-4o, a multimodal AI voice assistant that mixes textual content, imaginative and prescient, and audio, in April. The announcement put a brand new highlight on the prevailing “Sky” choice for ChatGPT’s Voice Mode, which some had likened to actor Scarlett Johansson’s fictional digital assistant within the 2013 movie Her — and sure, Johansson heard it too. The voice has been paused on ChatGPT for now.

In the meantime, in-company politics have reached boiling level for OpenAI, following a management disaster, a wave of excessive profile resignations and former executives calling out the corporate’s lack of transparency.