The stock of linguistic data that artificial intelligence like ChatGPT trains on could be exhausted by 2026. This is because AI consumes language data faster than it produces it.
technology
January 6, 2023
AI chatbots like ChatGPT may not do very well if they lack new, high-quality training data Ascanio/Alamy
Within three years, the supply of high-quality linguistic data used to train machine learning artificial intelligence models could run out, stalling AI progress.
Machine learning powers AI programs such as text prompt image generator Midjourney and OpenAI’s chat-based text generator ChatGPT. Such models train and learn from vast amounts of human-generated data from the internet. For example, when asked to draw a banana…