Imagine remembering a brief, wonderful romance with someone you had always admired. The conversations, the excitement, the feeling of being loved... Everything seems vivid and personally meaningful. Now imagine learning that a company could manufacture precisely that memory, including your inability to remember purchasing it. How much would the recollection mean to you once you began to question its origin?

Together with Andrea Lavazza and Iñigo de Miguel Beriain, I explore this question in our article, “The Pitt–Johansson Paradox: Artificial Memories and the Value of Authenticity,” published in Erkenntnis.  We examine a hypothetical world in which convincing artificial memories are available and awareness of that possibility becomes widespread. The puzzle concerns what happens to happiness when confidence in the origins of our memories becomes difficult to sustain.

The thought experiment that gives the paradox its name involves an implanted memory of a secret romance with Brad Pitt or Scarlett Johansson. These are purely illustrative examples. Secrecy is central to the scenario: a relationship remembered as private, ending amicably with a promise of discretion, would give the recipient reasons to avoid discussing it. A carefully constructed narrative could therefore limit encounters with evidence that might expose its artificial origin.

This type of memory could certainly be a source of great joy. We must consider this possibility carefully. For an individual whose life is marked by limited opportunities for satisfaction, a fabricated recollection may appear as a welcome means of finding solace. The philosophical dilemma becomes more complex when we acknowledge that memory manipulation might provide some of the advantages it promises.

The problem arises when individuals are aware that convincing memories can be fabricated, yet lack reliable methods to distinguish them from recollections of actual events. A person who holds dear an apparently genuine memory may start to question whether it was artificially implanted. This doubt could also impact someone whose cherished experience is genuinely rooted in reality. The technological possibility would have altered the context against which both individuals interpret their past.

Our formal analysis captures this effect through a straightforward decision-theoretic framework. We assume that an individual places greater value on a recollection believed to be authentic. As their confidence in its authenticity wanes, its expected value depreciates relative to the same recollection regarded as undoubtedly genuine. The magnitude of the loss hinges on the significance of authenticity to that person and the likelihood they assign to an artificial origin. This reveals a conditional cost that must be taken into account alongside any potential benefits of memory implantation.

The distinction between actual and perceived authenticity is crucial. Our argument deliberately leaves unresolved whether direct exposure to real events is essential for well-being. The model focuses on the contribution of perceived authenticity to the experiential value of an experience. It is possible that someone may derive pleasure from a memory they acknowledge as artificial; the mechanism we analyze is relevant when believing that an experience actually occurred enhances its value.

The consequences could also spill over into social life. Under the same circumstances, individuals might become less confident in others' recollections, thereby eroding testimonial trust. Even those who never purchase an artificial memory could still bear some of these costs. Reliable verification procedures could help mitigate the issue; our argument is most directly applicable in situations where verification is absent, imperfect, or inconsistently employed.

The Pitt–Johansson Paradox brings into focus an easily overlooked condition of remembered happiness: confidence in how a memory originated. If future technologies make desirable recollections more readily available, their impact will also depend on whether people can trust the connection between these memories and their actual experiences. What would be required to preserve that trust in a world where convincing memories could be manufactured at will?