Recently, OpenAI announced that the board of directors has established a security committee to oversee the direction of AI development, which is widely regarded as the long-term rumored development of GPT-5.

OpenAI stated in the announcement that the company has started training the next generation of cutting-edge models, and it is expected that the new system will take an important step towards artificial general intelligence (AGI).

Since the release of GPT-4 in March 2023, OpenAI has gradually launched a series of text, image, audio, and video generation applications centered around GPT-4. Just two weeks ago, the company released the latest version of GPT-4o. By significantly improving the interaction rate of the model, real-time interaction similar to human dialogue has been achieved. Despite improvements in multiple aspects, the GPT-4o still fails to surpass the capabilities of the GPT-4.

Therefore, the market has very high expectations for GPT-5. However, according to the latest statement from OpenAI, the next generation model may not be available to the public until next year. The training of AI models usually takes several months or even years, and the development team still needs several months to fine tune before they can be publicly released.

At the same time, as OpenAI's AI systems approach human level intelligence, the expression of AGI becomes more cautious. Anna Makanju, Vice President of Global Affairs at OpenAI, stated in an interview on Tuesday that OpenAI's mission is to build universal intelligence that can perform current human cognitive tasks.

A few years ago, Sam Altman confidently stated that the ultimate goal of OpenAI is to create super intelligence that is more advanced than humans. He also stated that he spends half of his time thinking about how to build super AI. However, Anna emphasized that the goal of OpenAI is to build AGI, not super intelligence, as the latter will be several orders of magnitude smarter than any human on Earth.

The latest announcement from OpenAI is also aimed at resolving another controversy - co founder Ilya Sutskever resigned earlier this month. He was responsible for the super alignment team that dealt with potential security issues in super intelligent systems. As the team disbanded, some team members criticized OpenAI management online for marginalizing the security team.

It is worth noting that the internal turmoil of OpenAI at the end of last year was also catalyzed by Sutskever, who played a key role in dismissing Altman. After Altman regained power, Sutskever did not hold a place on the new board of directors.

On Tuesday, OpenAI announced the establishment of a security committee led by Altman and three other directors, responsible for providing recommendations to the board on key security decisions related to OpenAI projects and operations. The main task of the new committee in the next 90 days is to evaluate and further improve the processes and security measures of OpenAI. Once approved by the board of directors, OpenAI will publicly share these updates.