Sam Altman, the CEO and co-founder of OpenAI, talked about the risks associated with AI systems. He said that OpenAI is not currently teaching GPT-5 and expected to be the next version of their AI language model.
On March 1, OpenAI changed its terms of service without announcing it. They said that they no longer use any data from APIs, which are tools that let people connect to OpenAI’s software.
Altman told at MIT event
Altman spoke at an event held at MIT where he was asked about an open letter that was circulated among the tech world. The letter requested that labs like OpenAI should stop developing AI systems that are more powerful than GPT-4. The reason behind the request was the concerns regarding the safety of future AI systems. However, the letter has faced criticism from many people in the industry, including some of the people who signed it. There is a disagreement among experts about how dangerous AI could be and how the industry could stop the development of such systems.
Altman, a person from MIT, said that the letter didn’t explain in enough detail when the pause is needed. The letter also said that OpenAI is training a new AI called GPT-5, but that’s not true. Altman said it’s a bit foolish to say that because it’s not happening anytime soon.
OpenAI’s new privacy
This update has occurred as industries are considering the possibility of large language models replacing content created by humans. For example, the Writers Guild of America is currently on strike after negotiations with movie studios fell through. The Guild has been pushing for limitations on the use of OpenAI’s ChatGPT for generating or rewriting scripts.
Although OpenAI is not currently developing GPT-5, they are still making improvements to GPT-4. OpenAI is also concerned about the safety implications of this work. According to Altman, they are working on other things related to GPT-4 that have important safety issues that were not mentioned in the letter.