According to the data head of Goldman Sachs, AI models such as chatgpt of Openai and Gemini of Google have been out of training data.
Nema Rafael, who serves as the Chief Data Officer of the banking giant and the head of data engineering, said that the issue can stop development artificial intelligence,
“We are already out of data,” Mr. Rafael said on the bank Exchanges Podcast, stating that the AI models are rapidly moving to the so -called synthetic data generated by artificial intelligence.
“I think what can be interesting. People may think that there may be a creative plateau … If all data is generated artificially, then how much can human data be included? I think it will be an interesting thing to look from a philosophical point of view.”
This is not the first time that senior industry data has expressed concern about the issue, which is “called”Peak“, AI models consume information about all the huge tribes of the Internet.
An article in the journal Nature In December, it was predicted that the “crisis point” would reach 2028. “Internet is a huge ocean of human knowledge, but it is not infinite,” Article Said. “Artificial intelligence researchers have almost dried it.”
Openai co-founder Ilya Sutsevar stated last year that the lack of training data would mean rapid development of AI “will” end unquestionably “.
According to Mr. Sutaskewar, the situation is similar to fossil fuels, as human-borne material is a finite resource like oil or coal.
“We have acquired peak data and no one else,” he said. “We have to deal with the data that we have. There is only one internet.”
Lack of new data AI can force companies to get away from current training models, more agents can focus on Artificial Intelligence with big language models such as Chatgpt.
The AI agents, which are already developed and released by most major artificial intelligence firms, work as autonomous systems that can make decisions and work online without human monitoring.