Artificial intelligence systems could devour all of the internet's free knowledge as soon as 2026, a new study has warned.
"If chatbots consume all of the available data, and there are no further advances in data efficiency, I would expect to see a relative stagnation in the field," study first author Pablo Villalobos, a researcher at the research institute Epoch AI, told Live Science."Models only improve slowly over time as new algorithmic insights are discovered and new data is naturally produced.
To estimate how much text is available online, the researchers used Google's web index, calculating that there were currently about 250 billion web pages containing 7,000 bytes of text per page. Then, they used follow-up analyses of internet protocol traffic — the flow of data across the web — and the activity of users online to project the growth of this available data stock.
The results revealed that high-quality information, taken from reliable sources, would be exhausted before 2032 at the latest — and that low-quality language data will be used up between 2030 and 2050. Image data, meanwhile, will be completely consumed between 2030 and 2060. "Companies are increasingly trying to use private data to train models, for example Meta's upcoming policy change," he added, in which the company announced it will use interactions with chatbots across its platforms to train its generative AI from June 26.