According to Goldman Sachs, artificial intelligence models such as Openai’s ChatGpt and Gemini Google have not used training data.
Neema Raphael, who serves as a chief banking giant and head of data engineering, said this could develop artificial intelligence.
“We’ve already finished the data,” Mr Rafael said of the bank. Exchange Podcasts added: Artificial intelligence models are increasingly converted into artificial data produced by artificial intelligence.
“I think what is interesting is that people may think there may be a creative plateau … If all data is produced artificially, then how much human data can be incorporated? I think this would be an interesting thing to watch from a philosophical perspective.”
This is not the first time that senior figures in the industry have raised concerns about the issue, which is referred to as “peak data”, where artificial intelligence models consume all the widespread Internet information.
An article in the journal Essence In December, he predicted that the “point of crisis” would be achieved by 2028. “The Internet is a broad ocean of human knowledge, but not infinite.” “AI researchers have almost dried it.”
Openai founder Ilya Sutour said last year that the lack of education data means that the rapid development of AI “will undoubtedly end.”
According to Mr Sutour, things are similar to fossil fuels, because human content is a limited source exactly like oil or coal.
“We have reached the peak of the data and there will be no longer there,” he said. “We have to deal with the data we have. There is only one internet.”
Lack of new data can force artificial intelligence companies to move away from current training models and change their focus from large language models such as ChatGpt to more artificial intelligence.
Artificial intelligence agents, currently developed and released by most important artificial information companies, serve as autonomous systems that can make decisions and perform their duties without human supervision.
