Understanding prevailing intelligence assessments can be difficult , given the swift evolution of AI technology . Several entities now generate scorecards that attempt to measure the capability of various AI models . These evaluations often consider several aspects, including correctness, speed,