This work explores the possibilities of utilizing machine learning to mitigate risks associated with the use of psychedelic substances. The problem is addressed through several classification models that predict the nature of a psychedelic experience.