AI has surpassed humans at a number of tasks and the rate at which humans are being surpassed at new tasks is increasing  

State-of-the-art AI performance on benchmarks, relative to human performance
Handwriting recognition  Speech recognition  Image recognition  Reading comprehension 
Language understanding  Common sense completion  Grade school math  Code generation 

For each benchmark, the maximally performing baseline reported in the benchmark paper is taken as the “starting point”, which is set at 0%. Human performance number is set at 100%. Handwriting recognition = MNIST, Language understanding = GLUE, Image recognition = ImageNet, Reading comprehension = SQuAD 1.1, Reading comprehension = SQuAD 2.0, Speech recognition = Switchboard, Grade school math = GSK8k, Common sense completion = HellaSwag, Code generation = HumanEval.
Chart: Will Henshall for TIME Source: ContextualAI
TIME