Authentication of encoded information is a popular current trend in optical security. Recent research has proposed the production of secure unclonable ID tags and devices with the use of nanoscale encoding and thin-film deposition fabrication techniques, which are nearly impossible to counterfeit but can be verified using optics and photonics instruments. Present procedures in optical encryption provide secure access to the information, and these techniques are improving daily. Nevertheless, a rightful recipient with access to the decryption key may not be able to validate the authenticity of the message. In other words, there is no simple way to check whether the information has been counterfeited. Metallic nanoparticles may be used in the fabrication process because they provide distinctive polarimetric signatures that can be used for validation. The data is encoded in the optical domain, which can be verified using physical properties with speckle analysis or ellipsometry. Signals obtained from fake and genuine samples are complex and can be difficult to distinguish. For this reason, machine-learning classification algorithms are required in order to determine the authenticity of the encoded data and verify the security of unclonable nanoparticle encoded or thin-film-based ID tags. In this paper, we review recent research on optical validation of messages, ID tags, and codes using nanostructures, thin films, and 3D optical codes. We analyze several case scenarios where optically encoded devices have to be authenticated. Validation requires the combined use of a variety of multi-disciplinary approaches in optical and statistical techniques, and for this reason, the first five sections of this paper are organized as a tutorial.