The Dark Side of Our Intuitions

Although research over the last century has significantly expanded our understanding of the problem-solving processes that precede such sudden solutions, and of the best strategies to elicit insight experiences, a less explored level of analysis concerns the role of the subjective phenomenology that accompanies these illuminating moments, the so-called “Aha!” moments.

The authors of the article The dark side of Eureka: Artificially induced Aha moments make facts feel true propose that eureka moments make an idea perceived as truer or more valuable, functioning similarly to a heuristic to facilitate rapid and efficient decision-making. Their central hypothesis is that facts, even if false, would appear more truthful if artificially accompanied by an “Aha!” moment elicited through an anagram task. To test this hypothesis, the authors conducted a pre-registered experiment with 300 participants. The results showed that participants provided higher truth ratings for statements accompanied by solved anagrams, even when the presented facts were false. This effect was particularly pronounced when participants reported having experienced an “Aha!” moment.

The article's introduction opens with the example of John Nash, the mathematician and Nobel laureate afflicted with schizophrenia, who explained his conviction of being recruited by aliens to save the world by stating that his ideas about supernatural beings came to him in the same way his mathematical ideas did, and for this reason, he took them seriously. This anecdote, although linked to a specific clinical condition, highlights a fundamental challenge for human beings in daily life: discerning between a true and useful idea and a false one, often in short timeframes to respond in a conversation, give advice, or solve a problem under pressure.

The authors question how the validity of a new idea is evaluated, especially when time is a critical factor. They suggest that, as hypothesized by Nash himself, individuals might rely on the phenomenology that accompanies their ideas, particularly “Aha!” moments. Currently, there is no universally accepted explanation for why the sensation of insight should predict accurate solutions to problems, but theoretical frameworks exist where this is not surprising.

It is plausible that the “Aha!” experience, like many other sensations that guide decision-making in productive ways, is a source of information for the problem-solver. Consequently, when a solution to a problem appears in the mind, the problem-solver can use the “Aha!” experience as a heuristic shortcut: a rapid evaluation of the idea's consistency with what they know, instead of engaging in a slow and laborious evaluation. The results of this study clearly demonstrate that an artificially induced "Aha! moment" can influence the judgment of a statement's truth. This result supports the idea that the sensation of insight can be used as a heuristic to evaluate the veracity of an idea, even when that sensation is generated by a source irrelevant to the statement's content. The implications of these results are significant. They suggest that our cognitive system can be deceptively influenced by our metacognitive experiences. At this level, I believe, lies the entire frontier dimension that characterizes our approach. Since an intense eureka sensation can make an idea convincing and true, regardless of its actual validity, this "dark side" of “Aha!” moments raises concerns in various contexts, including the formation of false beliefs, the spread of misinformation, and even the genesis of delusions in clinical contexts. This dark origin of eureka moments could be exploited by chatbots and language models in a dangerous manipulative path for people or to produce artificial “Aha!” moments. The example of John Nash becomes particularly relevant in this light, suggesting that the strong phenomenology associated with certain ideas (in that case, delusional ideas) can confer upon them an apparent intrinsic validity.

As always, we must ask ourselves how to harmonize human nature, with its characteristics that require attention and care, with the novelty of such pervasive and powerful machines that could produce unintended effects or, in the worst-case scenarios, could be used for social manipulations of various kinds.

This op-ed was originally published in Il Sole 24 Ore on June 4, 2025

Fr. Paolo Benanti is Visiting Distinguished Scholar at Seattle University. He is a force behind the multi-faith Rome Call for AI Ethics, a member of the UN High-level Advisory Board on AI Ethics and a professor of technology ethics, bioethics, artificial intelligence, and neuroethics at the Pontifical Gregorian University in Rome, Italy. He served as late Pope Francis's advisor on technology. He is the author of Homo Faber: The techno-Human condition, Digital age: Teoria del cambio d’epoca. Persona, famiglia e società, Noi e la macchina: Un'etica per l'era digitale, Ricordare troppo: Eccessi di memoria da Borges alle neuroscienze, Postumano, troppo postumano: Neurotecnologie e human enhancement, and Le macchine sapienti: Intelligenze artificiali e decisioni umane, among others. He contributes opinion pieces to Italian dailies regularly.