This paper consists of three parts. In the first part, I offer a conceptual analysis of the terms most commonly associated with falsehoods – misinformation, disinformation, and malinformation. I explore different ways in which this trichotomy has to be defined more precisely for us to understand the effects of each type of falsehood. In the second part, I provide a comprehensive definition of trust, which I then proceed to apply to different types of falsehoods. I outline four different sources of trust and try to match each to a complex taxonomy within what I call the ‘landscape of deception’. In the third part, I try to offer a broad suggestion for how standard approaches to solving the problem of trust in falsehoods could be improved in light of the preceding analysis. Finally, I try to show that an even greater concern potentially pertains to the appearance of different generative AI models.