OpenAI is training a new language model to replace GPT-4

Ever since the introduction of ChatGPT, all major companies have started developing their language models to compete with it.

OpenAI, creator of ChatGPT, recently stated that they are training a new language model, so let’s see what that means.

The new flagship language model is already in training

As the New York Times reports, OpenAI stated that it has begun training a new AI model that will replace GPT-4.

According to OpenAI, they expect the new model to bring the next level of capabilities while planning to build artificial general intelligence. The new model would act as an engine for various AI products.

The company is also creating a new Safety and Security Committee as Axios writes, that will assess the risk of this technology. OpenAI stated the following:

While we are proud to build and release models that are industry-leading on both capabilities and safety, we welcome a robust debate at this important moment.

The committee is advisory, and it will collaborate with an array of security experts, its first task is to enhance the OpenAI’s processes and safeguards.

It’s great to see that OpenAI plans to push the boundaries of artificial intelligence while seriously taking into consideration all the possible risks that come with it.

Do keep in mind that training an AI model can take months or years to complete. Once the training is finished, there’s a period of internal testing and tweaking before it can be released to the public.

While this announcement has many excited, it will be months before we can see this new model available to the public.

OpenAI CTO stated that a successor to GPT-4 will be unveiled later this year, so we can just sit tight and wait for the official announcement.

In case you missed it, OpenAI announced GPT-4o recently and it brings various improvements, including the ability to recognize voice input and hold conversations.

In other news, the new Copilot UI seems to take inspiration from ChatGPT, which doesn’t seem like a surprise since it already runs on GPT-4 technology.