Phosphoinositide 3-kinase (PI3K) and its product phosphatidylinositol(3,4,5)-trisphosphate (PIP3) control cell growth, migration, and other processes by recruiting proteins with pleckstrin homology (PH) domains and possibly other domains to the plasma membrane (PM). However, previous experimental and structural work with PH domains left conflicting evidence about which ones are PIP3 regulated. Here we used live-cell confocal imaging of 130 YFP-conjugated mouse PH domains and found that 20% translocated to the PM in response to receptor-generated PIP3 production. We developed a recursive-learning algorithm to predict PIP3 regulation of 1200 PH domains from different eukaryotes and validated that it accurately predicts PIP3 regulation. Strikingly, this algorithm showed that PIP3 regulation is specified by amino acids across the PH domain, not just the PIP3-binding pocket, and must have evolved several times independently from PIP3-insensitive ancestral PH domains. Finally, our algorithm and live-cell experiments provide a functional survey of PH domains in different species, showing that PI3K regulation increased from approximately two C. elegans and four Drosophila to 40 vertebrate proteins.