When artificial intelligence software like ChatGPT writes, it considers many options for each word, taking into account the response it has written so far and the question being asked. It assigns a score to each option on the list, which quantifies how likely the word is to come next, based on the vast amount of human-written text it has analyzed. ChatGPT, which is built on what is known as a large language model, then chooses a word with a high score, and moves on to the next one. The model's output is often so sophisticated that it can seem like the chatbot understands what it is saying -- but it does not. Every choice it makes is determined by complex math and huge amounts of data.

Two bird species, the Barbados bullfinch and the Carib grackle, have passed the string-pulling test, joining an elite group of animals capable of completing one of the most challenging animal cognition tests. The goal of the research, which was led by Jean-Nicolas Audet, a Ph.D. candidate in the Department of Biology at McGill University in Canada, is to determine the origin of innovative behavior in birds. Throughout the study, which was published Wednesday in the journal PLOS One, two of the 31 grackles and 18 of the 42 bullfinches were able to complete the string-pulling task, but what was surprising about their finding was lack of correlation between this and other cognitive tasks. "Most people previously assumed that string-pulling was just a problem-solving task and that the performance on both should be correlated. Some people even think that performance on all behavioral tasks should be correlated to some extent," Mr. Audet tells The Christian Science Monitor in an email.