The Text Dilemma: AI Struggles to Find Training Data for Chatbots

AI developers, including OpenAI, may be facing a shortage of text to train chatbots and large language models. Stuart Russell, an AI expert and professor at UC Berkeley, warned that the strategy of training AI models with vast amounts of text is "starting to hit a brick wall." Concerns have been raised about the data-collection practices of AI developers, including OpenAI, as they scrape public and private sources for text data. A study suggests that high-quality language data could be depleted by 2026. Lawsuits have also been filed against OpenAI, alleging the use of copyrighted materials and personal data in training their models.