Scanning probe microscopy (SPM) is a powerful technique for imaging surfaces with atomic precision, but it requires constant human supervision. This paper presents DeepSPM, an artificial intelligence framework for autonomous SPM operation. DeepSPM uses machine learning to autonomously select good sample regions, assess image quality, and condition the probe. It demonstrates continuous data acquisition in multi-day scanning tunneling microscopy experiments, managing probe quality in response to varying conditions. The approach enables advanced methods like large dataset acquisition and SPM-based nanolithography and can be generalized to most SPM techniques.

A. Krull, P. Hirsch, C. Rother, A. Schiffrin, C. Krull