AI doctors can perform on par with human doctors

On July 12, local time, researchers from Google and DeepMind published a study in the journal Nature, proposing a MultiMed QA assessment benchmark for evaluating the performance of big language models in terms of clinical knowledge. As a result, the researchers found that the AI doctors built by the big language model were comparable to human doctors in many aspects. Google's researchers said the technology would not threaten the work of general practitioners, but may be able to be used for medical helplines in the future. (Surge)