Using high throughput experimental data and in silico models to discover alternatives to toxic chromate corrosion inhibitors

•We screened a large library of organic compounds as replacements for toxic chromates.•High throughput automated corrosion testing was used to assess inhibitor performance.•Robust, predictive machine learning models of corrosion inhibition were developed.•Models indicated molecular features contributing to performance of organic inhibitors.•We also showed that quantum chemistry descriptors do not correlate with performance.