Laser interaction with solids is routinely used for functionalizing materials' surfaces. In most cases, the generation of patterns/structures is the key feature to endow materials with specific properties like hardening, superhydrophobicity, plasmonic colorâenhancement, or dedicated functions like antiâcounterfeiting tags. A way to generate random patterns, by means of generation of wrinkles on surfaces resulting from laser melting of amorphous Geâbased chalcogenide thin films, is presented. These patterns, similar to fingerprints, are modulations of the surface height by a few tens of nanometers with a subâmicrometer periodicity. It is shown that the patterns' spatial frequency depends on the melted layer thickness, which can be tuned by varying the impinging laser fluence. The randomness of these patterns makes them an excellent candidate for the generation of physical unclonable function tags (PUFâtags) for antiâcounterfeiting applications. Two specific ways are tested to identify the obtained PUFâtag: crossâcorrelation procedure or using a neural network. In both cases, it is demonstrated that the PUFâtag can be compared to a reference image (PUFâkey) and identified with a high recognition ratio on most real application conditions. This paves the way to straightforward nonâdeterministic PUFâtag generation dedicated to small sensitive parts such as, for example, electronic devices/components, jewelry, or watchmak.