The number of times a paper is cited is a poor proxy for its impact (see P. Stephan et al. Nature 544, 411–412; 2017). I suggest relying instead on a new metric that uses artificial intelligence (AI) to capture the subset of an author's or a paper's essential and therefore most highly influential citations.

Academics may cite papers for non-essential reasons — out of courtesy, for completeness or to promote their own publications. These superfluous citations can impede literature searches and exaggerate a paper's importance.

The scientific search engine, Semantic Scholar, is the first to automatically identify the subset of a paper's citations in which the paper had a strong impact on the citing work (see http://semanticscholar.org). It further ranks these according to their estimated impact by using machine-learning methods (see go.nature.com/2th2voa). Although still far from perfect, this 'highly influential citations' metric is a substantially better indicator of impact than raw citation counts are. An author's highly influential citation count is simply the sum of the highly influential citations of his or her papers.

This metric and its implementation exemplify the potential of AI to overcome information overload in the research literature.