On Tuesday local time, Open AI published a blog post saying that it has begun training a new generation of flagship artificial intelligence models to surpass the current GPT-4 model.

The company said the new model is expected to bring a higher level of capabilities and will be the core engine for various artificial intelligence products, including chatbots, intelligent assistants, search engines, and image generators, helping Open AI achieve the goals of “general artificial intelligence” (AGI).

There are still differences in the industry about when to achieve AGI, but technology companies such as Open AI, Google, Meta, and Microsoft have been steadily improving AI technology over the past 10 years, and will make a huge leap forward about every two to three years.

At the same time, Open AI also announced the formation of a new team of security advisors to address the risks that new models and future technology may present.

While we're proud to create and publish models that are industry-leading in terms of capabilities and security, we welcome intense discussions around this issue at this important time.

Open AI's goal is to advance AI technology faster while surpassing competitors, while also appeasing critics who think the technology is becoming increasingly dangerous, promotes the spread of disinformation, displaces jobs, and even threatens human survival.

GPT-4o, which was released earlier in May, demonstrated a more powerful ability not only to process text, but also to create image content, and answer questions and instructions in a highly conversational voice.

But it has also sparked some controversy, such as famous actress Scarlett Johansson's claim that GPT-4o uses a sound similar to her own.

Editor/Somer