Sam Altman, CEO and co-founder of OpenAI, has stated that the organization is not fully training the GPT-5 language model, which is intended to eventually replace the current GPT-4.
In answer to a question regarding a recent open letter signed by prominent members of the technology community, Altman made this revelation during a lecture at MIT. The letter called on AI companies, including OpenAI, to suspend training on platforms “more powerful than GPT-4” due to concerns about the security of future systems.
While Altman recognized the letter’s concerns, he said it “misses most of the technical nuances about where we need to pause.” He also stated that suspending work on GPT-5 does not mean the company has stopped improving GPT-4. Altman explained that OpenAI is currently focusing on other security issues that need to be addressed, which were not mentioned in the open letter.
It is important to note that GPT-5 development has just been postponed, not scrapped. However, Altman did not specify when the company intends to resume development on the neural network.
The OpenAI statement comes amid a rising discussion over the possible risks presented by AI algorithms and the necessity for businesses to secure the safety and security of their systems.
OpenAI’s decision not training the GPT-5 language model currently may alleviate some concerns raised by the open letter. However, AI companies must continue to prioritize security and address potential threats their systems pose. Only then can the benefits of AI be fully realized while minimizing the risks associated with this rapidly evolving technology.