The Dark Art of Dopamine Engineering

The Dark Art of Dopamine Engineering Jukka Ruohonen University of Southern Denmark Sønderborg, Denmark [email protected] Qusai Ramadan University of Southern Denmark Sønderborg, Denmark [email protected] Adam Alami University of Southern Denmark Sønderborg, Denmark [email protected] Abstract The paper presents a rapid review on what can be called—for a lack of a better term—dopamine engineering; the design of software, including but not limited to user interfaces, in a way that stimulates dopamine releases in the brains of the software’s users. According to the results, (i) the topic is interdisciplinary, covering not only computer science and software engineering but also neuroscience, psychology, medicine and life sciences in general, and law, among other disciplines, and the topics examined range from digital mental health and neuroethics to education, video games, and artificial intelligence. By and large, but not entirely, (ii) dopamine engineering has been discussed in relation to its negative consequences, including digital addiction. In addition, (iii) numerous factual statements are raised in the literature but evidence seems scant. Finally, (iv) law and ethics have occasionally been addressed but not in a systematic manner. The last two results justify a wording about dark art; not much is known about the topic according to the review results. CCS Concepts •Software and its engineering → Software design engineering. Keywords rapid review, software design, user interface design, human-computer interaction, cognition, psychology, manipulation, dark patterns ACM Reference Format: Jukka Ruohonen, Qusai Ramadan, and Adam Alami. 2025. The Dark Art of Dopamine Engineering. In .ACM, New York, NY, USA, 7 pages. https: //doi.org/10.1145/nnnnnnn.nnnnnnn 1 Introduction Dopamine has become a central metaphor in discussions of how digital technologies influence human behavior. In software engineering user interface design and social media platforms design choices are often described as ways of stimulating dopamine-driven reward mechanisms. Against this backdrop, the paper uses the term “dopamine engineering” not in a medical sense of altering neurotransmitters but in a sense of engineering digital human behavior that aims to capture and sustain attention. For instance, dark patterns—understood broadly as manipulative and (or) deceptive design choices [ 29 , 41 ], are supposedly often engineered with dopamine in mind. The so-called loot boxes used in many video games would be a good example in this regard [ 7 ]. Against this Conference’17, July 2017, Washington, DC, USA ©2025 Copyright held by the owner/author(s). Publication rights licensed to ACM. This is the author’s version of the work. It is posted here for your personal use. Not for redistribution. The definitive Version of Record was published in , https: //doi.org/10.1145/nnnnnnn.nnnnnnn. example and many others, it is no wonder that dopamine and its “engineering” are hotly debated all around the world, online and offline, in traditional media and in social media. Concepts such as dopamine fasting, dopamine detox, and digital detox have indeed long appeared in popular media [ 8 ]. They have also found their way to the scientific literature [ 13 ]. Elsewhere, developers have written memorandums for themselves and others on how to do dopamine engineering in practice [ 10 ]. However, no literature reviews have been written about what academic literature says about the topic. The absence of a review (even though rapid) is problematic because the debate around dopamine engineering is no longer confined to speculation or media hype; it is shaping public discourse, health practices, and regulatory agendas. For instance, many European countries have banned smartphones in elementary schools, some countries (such as Denmark, France, and Spain) have recently discussed banning social media for underage people, and some countries (notably, Australia) have already imposed such a ban upon children’s social media use [ 26 ]. At the time of writing, furthermore, the European Commission is seeking feedback on its proposal for a new Digital Fairness Act (DFA) seeking to counter manipulative and addictive digital designs [ 19 ]. With these policy responses in mind, the rapid review is a step to a consolidated understanding of how the scientific community has engaged with the concept. Moving from a fragmented understanding can inform future research directions and contribute to policy discussions. The paper and its remaining structure is straightforward. Thus, without further ado, Section 2 presents the research questions examined, Section 3 the reviewing materials and methods, Section 4 the results, and Section 5 the conclusion, including a brief discussion about what the literature reviewed indicates for further research. 2 Research Questions Four research questions (RQs) are examined: • RQ.1: In addition to either explicit or implicit framing toward software engineering, in which scientific fields dopamine appears in the literature and what topics are addressed? •RQ.2: How is dopamine contextualized in the literature? • RQ.3: What factual statements, including but not limited to empirical evidence, are made in the literature about dopamine? •RQ.4: Are ethics and law discussed and how? Of these research questions, RQ.3 requires a clarification. By the term factual statements, we mean strong or even definite statements about some facts related to dopamine, regardless whether there is evidence backing them up. Thus, the goal of the review is not to evaluate evidence—if there is any—but to summarize factual claims raised, regardless whether they are true, false, or somewhere inbetween. Even then, already the summary can be used to tentatively probe whether there are also so-called “Leprechaun claims” in the literature, as has been common in software engineering [6]. Conference’17, July 2017, Washington, DC, USA Jukka Ruohonen, Qusai Ramadan, and Adam Alami 3 Materials and Methods The paper is a rapid review of software engineering literature on dopamine. Much like scoping reviews, rapid reviews use the techniques of systematic literature reviews (SLRs) but with feasibility in mind [ 40 ]. Typically, the search protocols used in SLRs are altered in rapid reviews to keep the amount of literature feasible for reviewing under time and other resource constraints. The literature search is based on Scopus. When executed on 21 September 2025 from papers’ all fields, search strings “dopamine” and “software engineering” resulted 407 papers. The amount is infeasible for the present purposes. When skimming through the papers’ titles, furthermore, most of them had nothing to do with software engineering as a discipline or a practice. The publication venues were also in biomedical and related disciplines. By adding a third search string “user interface”, only 30 papers were returned. These were included in the initial sample for screening. Further 13 papers were returned by replacing the third search string with “software design”. These were also added to the screening sample. As seen from Fig. 1, a further replacement with a string “mobile application” resulted 16 papers. Social media brought additional 39 papers. Yet another try with “dark pattern” as a third search term resulted only three papers but they were all duplicates with the previous queries. Figure 1: Literature Sampling Regarding qualification criteria into the final reviewing sample, (a) nine duplicates were removed, (b) 11 papers or books had to be excluded due to a lack of access, (c) 30 papers were excluded because the noun dopamine only appeared in the titles of papers referenced by a given paper in the screening sample without a notable traceable argument raised in the paper, (d) 15 papers were excluded because there were not about software engineering after all, (e) four papers had to be excluded due to errors in Scopus, (f) one paper was excluded because it was not about dopamine but an artifact named as such, and, finally, (g) one paper was excluded due to likely hallucinated (incorrect) references [ 46 ]. With respect to the errors, which have been observed also previously [ 20 , 40 ], some of the keywords did not actually appear in two papers and no traces were available in the Internet about two papers according to Google searches, including with the papers’ digital object identifiers. Thus, in total, the sample rapidly reviewed contains 𝑛= 28 papers. By following guidelines [ 25 , 30 ], the results are presented as a narrative and tabular synthesis initially done by the first author and rapidly audited by the remaining authors. In practice, the reviewing authors were instructed to search for paragraphs containing the nouns dopamine (RQ.2 and RQ.3) and ethics (RQ.4). Given the narrative focus and the RQs specified, inter-rater reliability measures are not needed because none of the research questions are about evidence per se. In addition, a qualitative, human-made categorization is presented only for summarizing the answer to RQ.3. In contrast, the categorization for RQ.1 is done by using OpenAlex. 1 Although hardly perfect, as none of the scientific database tools are, it provides a scientific field, a scientific sub-field, and a topic for each paper, book, or other material in its database. Particularly the fields and sub-fields have been seen as useful for high-level mappings of scientific topics [ 21 , 33 ]. Given RQ.1 and the keywords in Fig. 1, interdisciplinarity can also be expected almost by definition. 4 Results In what follow, the reviewing results are presented by following the ordering of the RQs. For presentation reasons—specifically to avoid incorrect referencing, citations are omitted from the quotes presented. The same applies to type-setting techniques, such as italics and boldfaced text, which were removed from quotations. 4.1 Fields and Topics (RQ.1) The scientific fields and sub-fields identified by OpenAlex via its machine learning algorithms are enumerated in Table 1. The first thing to note is that software engineering does not appear in the table, but this result may well be because OpenAlex does not recognize it as a distinct discipline. Nevertheless, computer science is well-represented, so the question is more about semantics and disciplinary issues. The second thing to notice is a good representation of psychology and cognitive sciences as well. Also social sciences, including everything from law to political science, have contributed to contemplations about dopamine in the software engineering context. Thus, the topic is interdisciplinary, as expected. When looking at the topics identified by OpenAlex in Table 2, the variance is large, including everything from digital mental health issues to emotion recognition via artificial intelligence (AI) algorithms, privacy and security, and even smart cities. However, it should be noted that a qualitative categorization would have resulted a different and arguably a better summary. For instance, the paper [ 37 ] could be better categorized into the AI-driven emotion detection and gamification topics. Another point is that some notable topics are a little surprisingly missing. For instance, recommender systems do not appear in the table, although these systems have been seen as particularly dopamine-driven and manipulative by many users [ 29 , 39 ]. Despite the search strings used, OpenAlex also categorizes only one paper to be about social media. For these reasons, a narrative summary about the contextualizations presented in the papers reviewed provide a more refined and nuanced picture. 4.2 Contextualizations (RQ.2) The remark in the introduction about intentionally engineering software to cause dopamine bursts is recognized in the literature. In fact, not only is dopamine noted but it is asserted that some companies are intentionally making their software products addictive: 1https://openalex.org/ The Dark Art of Dopamine Engineering Conference’17, July 2017, Washington, DC, USA Table 1: Fields and Sub-fields According to OpenAlex Field Papers Sub-fields Papers •Computer science1[1, 2, 17, 34, 43, 44, 48, 49] •Artificial intelligence [44, 48, 49] •Computer science applications [1] •Computer vision and pattern recognition [34] •Human-computer interaction [43] •Information systems1[2, 17] •Decision sciences [31] •Information systems and management [31] •Economics, econometrics, and finance [24] •General economics, econometrics, and finance [24] •Engineering [37, 38] •Media technology [37, 38] •Medicine [16] •Neurology [16] •Neuroscience [12, 35] •Cognitive neuroscience [12, 35] •Physics and astronomy [28] •Statistical and nonlinear physics [28] •Psychology [3, 4, 11, 14, 18, 22, 45] •Applied psychology [22, 45] •Development and education psychology [11, 18] •Experimental and cognitive psychology [4] •Social psychology [3, 14] •Social sciences [9, 15, 23, 36, 42] •Health [9] •Law [15] •Sociology and political science [23, 36, 42] 1The paper [2] is not present in OpenAlex; it was added manually to these categories. Table 2: Topics According to OpenAlex Topic Paper •Behavioral health interventions [22] •Complex network analysis techniques [28] •Digital mental health interventions [45] •Economic development and digital transformation [24] •Emotion and mood recognition [4] •Reinforcement learning in robotics [49] •Educational games and gamification1[2, 11, 18] •Embodied and extended cognition [12] •Hate speech and cyber-bullying detection [44] •Human-automated interaction and safety [14] •Impact of technology on adolescents [23] •Innovative human-technology interaction [43] •Intelligent tutoring systems and adaptive learning [48] •Law in society and culture [15] •Multi-modal machine learning applications [34] •Neuroethics and biomedical innovations2[35] •Open source software innovations [1] •Parkinson’s disease mechanisms and treatments [16] •Personal inf. management and user behavior [31] •Privacy, security, and data protection [36, 42] •Psychological well-being and life satisfaction [3] •Scientific research and philosophical inquiry [17] •Smart cities and technologies [37, 38] •Social media in health education [9] 1The paper [2] is not present in OpenAlex; it was added manually to this category. 2Also “human enhancement” belongs to this topic, but it was omitted due to a lack of space. (1) “But why is this so? Are all social networks and entertainment applications addictive? Of course not, but the most popular ones, such as Facebook, have entire departments that deal with human psychology and create more interactive user interfaces, making applications and websites addictive.” [31, p. 4] The above quotation (1) is a good example about reservations regarding factual statements; not everything is about dopamine engineering, and this truism is good to keep in mind also regarding what is subsequently presented. Another point is about the naming of a particular social media platform and its interdisciplinary engineering. While Meta, Facebook, and social media in general are well-represented also in popular and political discourses, there is a slightly more illuminating example presented in the literature: (2) “Take, for example, Dopamine Labs (also known as Boundless Mind: www.boundless.ai/), a company that sells neurobiologically tuned algorithms for manipulating users to the application developers.” [35, p. 141] The quotation (2) asserts that dopamine engineering in some parts of the software industry involves also consulting, softwareas-a-service solutions, and generally outsourcing of addiction generation solutions. The company mentioned in the quote indeed allegedly “provides services to help application developers design addicting applications” [ 35 , p. 114]. It is worth to continue the quote a little further because the subsequent sentences motivate the points about ethical and legal issues discussed later in Subsection 4.4. Thus: (3) “Dopamine Labs and its clients are not concerned with whether you develop bad habits or if their technologies lead to problematic uses of the Internet, online gambling, smartphones, and video games. Strangely enough, the company is open to making money off persons who they have helped to addict.” [35, p. 141] Conference’17, July 2017, Washington, DC, USA Jukka Ruohonen, Qusai Ramadan, and Adam Alami Table 3: Factual Statements About Dopamine in the Literature •Digital addiction may affect the brain’s reward systems, impacting emotion regulation [23] • Dopamine is associated with movement, attention, and memory [ 18 , 34 , 37 , 45 ], thus, among other things, mediating the motivation and cognitive effort required for task execution and learning [1, 11, 14, 34, 37], and being also associated with Parkinson’s disease [16, 45] •Hate speech and negative comments cause addiction due to dopamine releases, further causing psychological issues [44] •Mobile applications cause instant gratification through over-release of dopamine, further causing emotional issues [17] •Problematic Internet use damages the brain’s dopaminergic system [12], causing antisocial behavior [45] • Some neural network algorithms, including reinforcement learning in particular, work analogously to the release of dopamine [ 1 , 4 , 48 , 49 ] •Social media is not only associated with dopamine [28] but also adrenaline and cortisol levels [9, 28] • Social media may cause [ 9 , 24 ] or causes addiction due to intentional dopamine engineering, including in relation to feeds, notifications, likes, comments, and other algorithmic and user interface designs [17, 31, 35, 42, 45] •Social media or smartphones cause addiction among children due to dopamine [43] •Stress can dysregulate dopamine levels, causing mood disorders [36] •Stress is not negatively associated with mood while using mobile dating applications [36] •There are some companies doing intentional dopamine engineering to get users addicted [12, 17, 31, 35, 42, 45] •Video games have the potential to elevate dopamine levels [11, 37, 38] •Video games increase dopamine levels during playing and the effect is sustained after gaming is over [22] •Video games increase dopamine levels to an extent comparable to taking amphetamines [2, 3] •Video games stimulate dopamine releases and thus cause addiction [35] Thus, for the author, dopamine engineering more or less equates to selling of drugs. And, indeed, in the popular online engineering discourse even hard drugs such as cocaine are mentioned alongside user interface and software design in general [ 10 ]. Opiates and amphetamines are mentioned also by the author of the above quote [ 35 ]. Getting addicted to cocaine, opiates, or amphetamines has obviously consequences for individuals and societies, but what about digital addiction? For one thing, there is the widely debated diminution of concentration, which is allegedly caused by instantly available pleasure, which, in turn, is caused by the excessive dopamine releases that supposedly have emotional consequences: (4) “Gone are the days when one had to visit a friend’s house to inquire if they wished to go out; now we simply ask via WhatsApp and receive a response within seconds. We live in a society of instant gratification, which leads to an overactivation of the dopaminergic system and emotional deregulation.” [35, p. 7] However, it is not only mobile applications and social media providing the instant pleasures causing addictions. Allegedly, video game playing has “also been found to result in substantial dopamine release in the dopaminergic system as well as addiction” [ 35 , p. 134]. This statement about video games is a good motivation for the results presented in the subsequent Subsection 4.3. The quote (4) originates from a broader theorization considering not only neurology but also social sciences, among them social psychology: (5) “For example, connecting with others socially is a deeply rooted desire in humans and our brains have evolved to release the dopamine reward every time we use these technologies.” [35, p. 140] However, not all allegedly enjoy jolly social meetings for raising their dopamine levels even though social behavior is still present: (6) “Dopamine Addiction: Nasty comments generate a kind of addiction in individuals, as they prefer to read the negative comments repeatedly when bored or for pleasure at times, causing a mental health issue and negativity in the persona, leaving the person uninspired.” [44, p. 7548] Above, again, the keyword is addiction from which the consequences allegedly follow. By implication, the quote (6) is related to the earlier quotes (1), (2), and (3) in a sense that software engineering addictive social media platforms likely facilitates the behavior described. Digital addiction indeed is the underlying theme in many of the papers. To this end, some have even warned their fellow scientists about the dangers involved with social media: (7) “One should remember that tweeting can be addictive and should therefore be self-managed with care. Getting notifications on social media activates the same dopamine loop as gambling. When you become addicted to Twitter and you are not able to check your phone, adrenaline and cortisol are released to your body.” [9, p. 5] This statement is a good example because it also illustrates what might called anti-dopamine effects arising from not being able to use social media to get one’s dose of dopamine. In addition, the naming of a particular, now renamed social media platform reiterates the earlier points about dopamine engineering. In addition, the context of science motivates the following further quotation that connects anti-dopamine effects to teaching and learning in higher education: (8) “Also, they could experience anxiety and demotivation due to competition associated teaching and learning methods, triggering an anti-dopamine response and dissatisfaction. As a counter response, students might engage in increased and even late-night interactions on social networks, contributing to insomnia related effects.” [24, p. 194] Above, the message is that some pedagogical choices may fail to deliver dopamine needed for learning, which students allegedly then seek elsewhere, including from social media. However, the quote (8) is entirely speculative. Therefore, regarding the factual The Dark Art of Dopamine Engineering Conference’17, July 2017, Washington, DC, USA statements considered next, the earlier quote (7) is a better example due to the wording that social media “can be addictive”. In other words—and unlike with many other authors of the papers reviewed, the author does make a definite statement about social media strictly causing addiction—always, everywhere, and for everyone. 4.3 Factual Statements (RQ.3) The previous quotations already contain many factual statements, which, to recall, may or may not be true—and the evaluation of their truthfulness and truthlikeness [ 32 ] is not the rapid review’s goal. With this reiterating point in mind, Table 3 summarizes the remaining factual statements. Given the context of higher education to which the previous Subsection 4.2 ended, it is worth starting from the second row in the table according to which dopamine is also associated with many factors essential for human flourishing, including motivation, attention, and stimulation of cognitive tasks. Although dopamine engineering in a sense described earlier, including with references to hard drugs, is well-represented in Table 3, it should be thus underlined that some engineering associated with dopamine is also sinister and beneficial. For instance, the sample includes also a paper that designed a software solution for improving drug delivery for countering a dopamine withdrawal symptom associated with the Parkinson’s disease [ 16 ]. That said, the clear majority of the factual statements raised in the literature are about the negative consequences from dopamine engineering. The noun engineering is further reinforced by the computer science studies that have contemplated dopamine, or having been inspired by it, when designing neural networks and reinforcement learning algorithms in particular. And there is only a small step from reinforcement learning to the earlier quotes (1), (2), and (3). The seemingly general statements about dopamine on the second line of the table have been used to motivate a video game design in relation to suspense, motivation, and surprise [ 18 ]. In the online guidelines for dopamine engineering their addictive potential is a driving motivation [ 10 ]. Regarding the statement about problematic Internet use, the quotes (1), (2), and (3) again apply indirectly. Regarding the earlier points about semantics involved when raising factual statements, the three lines in Table 3 about social media are a good example because some authors are much more cautions than others. To put aside the study about children, which seems a special and particularly sensitive case, some authors have only asserted that social media may cause addiction due to dopamineseeking, whereas others have been sure due to definite claims put forward. Furthermore, the wording about dark art in the paper’s title can be justified by the widespread but speculative listing of particularly design choices causing addiction. Feeds, likes, and related things have been frequently mentioned but the literature is unable to conclude which ones are particularly risky. In this regard, some have asserted that likes and alike are not a sufficient condition for establishing addiction [ 35 , p. 139], instead arguing along the lines of the earlier quote (7). Video games are also a good example. When reading the last four lines in Table 3, the varying assertion and speculation levels are again visible. Some have written that video game playing may increase dopamine levels. Others have asserted that playing increases the levels. Both citing an old paper [ 27 ], a couple of papers have claimed that the dopamine levels gained from playing are comparable to those caused by amphetamines. The last line in the table simply asserts that video game playing causes addiction. Regardless whichever statement is true or false, or more probable than some other, the wording about dark art is again manifested because the papers reviewed cannot agree—and at the same time truthlikeness is asserted regarding dopamine engineering as an industry practice comparable to drug dealing. 4.4 Ethics and Law (RQ.4) By using a similar criteria that was used to collect the literature sample (see Section 3), such that mentions in references and ethical study approvals are excluded, the nouns ethics or ethical appear only in about 29% of the papers and the nouns law or legal in about 14% of them. In addition, some papers merely mention these in footnote-like sentences. For instance, one paper merely noted ethics and international law in relation to military contexts [ 49 ]. Another paper likewise noted the importance of ethical consciousness in designing of virtual worlds [ 34 ], but without going deeper than a note. That said, particularly the few monographs in the reviewing sample have addressed ethics to a great extent. For instance, one of these derived six ethical principles to guide an ethical design of software that may have negative consequences for digital behavior [ 17 , p. 188]. Also the ethics of both social media and research of social media were thoroughly addressed in another monograph [ 35 ]. The latter were also noted in a paper about so-called troll factories [ 28 ]. Also the ethics of video games were addressed. Drawing on Kantian deontological ethics, the author of a monograph concluded that as “long as players treat each other with respect, playing video games does not result in unambiguous harm” [ 35 , p. 268]. In addition, for “the utilitarian, several claims about the potential harms and negative consequences of video game play are negated due to the robust empirical evidence of benefits that outweigh the less substantiated shortcomings” [ 35 , p. 268]. The earlier quote (6) could be used as a counterargument to the Kantian take, whereas the truthlikeness of the factual claim about “robust empirical evidence” would warrant a careful assessment. The alleged benefits of gaming should be evaluated also against the other factual statements listed in Table 3. More generally, interdisciplinary approaches were emphasized: (9) “This perspective underscores the need for a more comprehensive framework that integrates cognitive science, political philosophy, and critical theory to fully grasp the techno-political dynamics of attention capture and its implications for human agency in the age of cognitive capitalism.” [12, p. 11] In addition to interdisciplinarity, human agency is a central keyword in the above quote (9). Ethical concerns related to agency and autonomy were also emphasized in a paper dealing with a datafication and surveillance of children and their childhoods [ 43 ]. The following further quote summarizes the fundamental issue well: (10) “The extent to which such approaches have moved beyond the ‘libertarian paternalism’ of so-called ‘nudging’ (intended as it was to strike a balance in the civic sphere between the individual’s freedom to choose and ‘better’ societal) and into the realm of manipulation and even the cultivation of ‘tech addiction’ is an emerging topic in both the academy and civil society.” [ 15 , p. 61] Conference’17, July 2017, Washington, DC, USA Jukka Ruohonen, Qusai Ramadan, and Adam Alami Similar points have been raised in the non-sampled literature on dark patterns and their relation to the difficult concepts of deception and manipulation [ 41 ]. Finally, it is worth remarking that the literature recognized the practical problems involved with applied ethics. Accordingly, ethics should not be left as a responsibility of individual designers and engineers—they should be rather recognized at the top-levels of organizations and their stakeholders: (11) “It would be simplistic to suggest that if a company refuses to enforce ethical standards, the digital behavior designer should resign”, and, therefore, “in everyday practice, ethical consideration in the design and development of digital services is infrequent”, implying that ethics “must be acknowledged and valued by stakeholders as of paramount importance, ensuring that ethical decisions are adopted uniformly, from the highest corporate levels to the design and development stages” [17, pp. 188–189, 193]. All in all, ethics have been recognized in many of the papers reviewed, but not in all papers and not systematically in general. Furthermore, law has received a limited attention in the literature in general, although there is a monograph about it [ 15 ]. Otherwise, legal aspects have mostly appeared as short remarks. For instance, after elaborating and discussing “economic and psychological dangers and risks of user addiction to social networks”, the authors of a paper explicitly called for regulatory interventions [ 42 , p. 276–278]. Another paper merely noted that legal responsibilities should be taken into account in software design [ 38 ]. The same applies to data protection and privacy concerns [ 43 ]. Given the close relation between deontological ethics and law, these points reinforce the argument about non-systematic assessments about law and ethics. 5 Conclusion The paper investigated what was called “dopamine engineering”, the engineering of software explicitly with dopamine in mind. Four research questions were examined. Regarding RQ.1, the answer is clear: the topic is interdisciplinary, covering not only software engineering and computer science but also neuroscience, medicine, psychology, law, and many other branches of social sciences, among other disciplines. However, the disciplines involved and the topics addressed are both very heterogenous, suggesting that dopamine has not yet been systematically explored. As for RQ.2, dopamine and its “engineering” have largely been contextualized by considering their negative consequences, including particularly digital addiction and certain companies driving it through their design choices and algorithms. However, none of the papers reviewed propose tools or systematic methodologies for detecting, evaluating, or avoiding dopamine-related design choices during a software’s development life cycle. Nor are risks such designs may pose properly addressed. These conclusions align with the subsequent answer to RQ.3. Thus, with respect to RQ.3, numerous factual statements but with a little empirical evidence backing them have been made in the literature. Many of these are related to digital addiction. As some acknowledged in relation to a claim about video games, “there is need for further replication” [ 35 , p. 134]. This good point applies to most factual statements identified in Table 3. In fact, many of the statements in the table come with no evidence whatsoever, implying that the earlier remark about “Leprechaun claims” in Section 2 can be argued to apply in overall. Such claims are particularly dangerous in the present context because of their relation to public discourse, health practice, and regulatory agendas noted in the introduction. Although the notion of evidence-based policy-making has long been criticized [ 5 , 47 ], academics should arguably avoid contributing to such policy-making in case their evidence is only tentative at best. Of course, they can contribute like anyone in a democracy but the noun evidence-based should be reserved for cases whose truthfulness and truthlikeness can be assessed to a reasonable degree. This point is strengthened by an overall lack of research on the already existing legislation affecting the topic. Legislation also explains why RQ.3 is not only empirical evidence. As for RQ.4 more generally, ethics have been addressed better than law but still largely in a non-systematic manner. Of the philosophical concepts involved, concepts such as individual autonomy and human agency,manipulation,deception,nudging, and freedom of choice have been used to ground the ethical assessments. Given the fundamental nature of the concepts, interdisciplinary approaches have been recommended also regarding ethics. Thus: in addition to empirical validation and replication research regarding the factual statements expressed, more theoretical, legal, and philosophical research is required too for better understanding the topic of dopamine engineering. Eventually, ethical and theoretically grounded evidence could also influence policy-making. A similar conclusion applies to the wording in the paper’s title about “dark art”. That is, the notion of “dopamine engineering”, whether in terms of user interface design or reinforcement learning of user behavior, is arguably without much robust empirical evidence. References [1] V. Anand and C. Hota. Motivation of Participants in Crowdsourcing Platforms Using Intelligent Agents. International Journal of Computing, 19(1):78–87, 2020. [2] L. Argenton, F. Pallavicini, G. Barone, A. Zini, G. Garcea, A. Zanacchi, and F. Mantovani. Hacking Well-Being: Can Serious Games Fit Positive Technology Framework? In D. Villani, P. Cipresso, A. Gaggioli, and G. Riva, editors, Integrating Technology in Positive Psychology Practice. IGI Global, London, 2016. [3] L. Argenton, E. Schek, and F. Mantovani. Serious Games as Positive Technologies. In Proceedings of the 6th International Conference on Virtual, Augmented and Mixed Reality: Applications of Virtual and Augmented Reality (VAMR 2014), pages 169–177, Heraklion, 2014. Springer. [4] G. Assunção, B. Patrão, M. Castelo-Branco, , and P. Menezes. An Overview of Emotion in Artificial Intelligence. IEEE Transactions on Artificial Intelligence, 3(6):867–886, 2022. [5] H. Berg. Is Evidence-Based Practice Justified?—A Philosophical Critique. Journal of Evaluation in Clinical Practice, 30:855–859, 2024. [6] L. Bossavit. The Leprechauns of Software Engineering. Leanpub, Victoria, 2015. [7] A. Brady. Are Loot Boxes Addictive? Analyzing Participant’s Physiological Arousal While Opening a Loot Box. Games and Culture, 16(4):419–433, 2019. [8] O. Burkeman. Dopamine Fasting is a Tech Fad That Sounds silly – But Might Just Work. The Guardian, available online in September: https://www.theguardian. com/lifeandstyle/2019/nov/29/dopamine-fasting-tech-fad-might-work, 2019. [9] V. Cheplygina, F. Hermans, C. Albers, N. Bielczyk, and I. Smeets. Ten Simple Rules for Getting Started on Twitter as a Scientist. PLOS Computational Biology, 16(2):1–9, 2020. [10] P. Chopra. A Primer on Dopamine. Inverted Passion Blog, available online in September 2025: https://invertedpassion.com/a-primer-on-dopamine/, 2024. [11] F. Coelho, D. Gonçalves, and A. M. Abreu. The Impact of Game-Based Interventions on Adult Cognition: A Systematic Review. Journal of Human–Computer Interaction, pages 1–36, 2025. [12] P. G. de la Torre, M. Pérez - Verdugo, and X. E. Barandiaran. Attention Is All They Need: Cognitive Science and the (Techno)Political Economy of Attention in Humans and Machines. AI & SOCIETY, pages 1–7, 2025. [13] D. Desai, J. Patel, F. Saiyed, H. Upadhyay, P. Kariya, and J. Patel. A Literature Review on Holistic Well-Being and Dopamine Fasting: An Integrated Approach. Cureus, 16(6):e61643, 2024. The Dark Art of Dopamine Engineering Conference’17, July 2017, Washington, DC, USA [14] M. Diarra, J. Theurel, and B. Paty. Systematic Review of Neurophysiological Assessment Techniques and Metrics for Mental Workload Evaluation in RealWorld Settings. Frontiers in Neuroergonomics, 6:1–19, 2025. [15] L. E. Diver. Digisprudence: Code as Law Rebooted. Edinburgh University Press, Edinburgh, 2022. [16] D. N. Domen Verber and, M. Borovič, J. Dugonik, and D. Flisar. EQUIDopa: A Responsive Web Application for the Levodopa Equivalent Dose Calculator. Computer Methods and Programs in Biomedicine, 196:105633, 2020. [17] U. L. Domínguez. DIGITAL BEHAVIOR: Evidence-Based Digital Design. Cambridge University Press, Cambridge, 2025. [18] P. DSouza, D. F. Galeote, N.-Z. Legaki, and J. Hamari. Edutainment in the Context of Science Communication in Virtual Reality. In Companion Proceedings of the 2024 Annual Symposium on Computer-Human Interaction in Play (CHI PLAY Companion 2024), pages 359–364, Tampere, 2024. ACM. [19] European Commission. Commission Launches Open Consultation on the Forthcoming Digital Fairness Act. Available online in September: https://digital-strategy.ec.europa.eu/en/consultations/commission-launchesopen-consultation-forthcoming-digital-fairness-act, 2025. [20] F. Franceschini, D. Maisano, and L. Mastrogiacomo. The Museum of Errors/Horrors in Scopus. Journal of Informetrics, 10(1):174–182, 2016. [21] R. Haunschild and L. Bornmann. The Use of OpenAlex to Produce Meaningful Bibliometric Global Overlay Maps of Science on the Individual, Institutional, and National Levels. PLOS ONE, 19:e0308041, 2024. [22] P. Jerčić and V. Sundstedt. Practicing Emotion-Regulation Through Biofeedback on the Decision-Making Performance in the Context of Serious Games: A Systematic Rview. Entertainment Computing, 29:75–86, 2019. [23] T. Karakose, B. Yildirim, T. Tülübaş, and A. Kardas. A Comprehensive Review on Emerging Trends in the Dynamic Evolution of Digital Addiction and Depression. Frontiers in Psychology, 14:1–24, 2023. [24] R. Kazakov, P. Petrova, and Y. Kazakova. Understanding and Managing Socioeconomic Systems Behaviour. Springer, Cham, 2024. [25] V. J. King, B. Nussbaumer-Streit, E. Shaw, D. Devane, L. Kahwati, M. Viswanathan, and G. G. on behalf of the Cochrane Rapid Reviews Methods Group. Rapid Reviews Methods Series: Considerations and Recommendations for Evidence Synthesis in Rapid Reviews. BMJ Evidence-Based Medicine, 29(6):419–422, 2024. [26] P. Kirby. Ban Social Media for Under-15s, Says French Report Warning of TikTok Risks. BBC, available online in September: https://www.bbc.com/news/articles/ crkjep23403o, 2025. [27] M. J. Koepp, R. N. Gunn, A. D. Lawrence, V. J. Cunningham, A. Dagher, T. Jones, D. J. Brooks, C. J. Bench, and P. M. Grasby. Evidence for Striatal Dopamine Release During a Video Game. Nature, 393:266–268, 1998. [28] F. S. Marcondes, J. J. Almeida, and P. Novais. An Exploratory Design Science Research on Troll Factories. Integrated Computer-Aided Engineering, 31:95–116, 2024. [29] C. Mohanlal. Investigation: Platforms Still Use Manipulative Design Despite DSA Rules. DSA Observatory, available online in September: https://dsa-observatory.eu/2025/08/07/investigation-platforms-still-usemanipulative-design-despite-dsa-rules/, 2025. [30] P. Moons, E. Goossens, , and D. R. Thompson. Rapid Reviews: The Pros and Cons of an Accelerated Revie Process. European Journal of Cardiovascular Nursing, 20:515–519, 2021. [31] Y. Neznaradko and L. Demkiv. Mobile Application of Information System for Smartphone Use Control. In Proceedings of the 3rd International Workshop on Modern Machine Learning Technologies and Data Science, pages 1–47, Lviv-Shatsk, 2021. CEUR-WS. [32] I. Niiniluoto. Critical Scientific Realism. Oxford University Press, Oxford, 2002. [33] K. Okamura. Evolving Interdisciplinary Contributions to Global Societal Challenges: A 50-Year Overview. Archived manuscript, available online: https: //doi.org/10.48550/arXiv.2410.20619, 2024. [34] S.-M. Park and Y.-G. Kim. A Metaverse: Taxonomy, Components, Applications, and Open Challenges. IEEE Access, 10:4209–4251, 2022. [35] T. D. Parsons. Ethical Challenges in Digital Psychology and Cyberpsychology. Cambridge University Press, Cambridge, 2019. [36] L. Perić, M. Rabar, and T. Orehovački. The Impact of Dating Apps on the Mental Health of the LGBTIQA+ Population. Multimodal Technologies and Interaction, 9:1–40, 2025. [37] M. Rodrigues, R. Machado, R. Costa, and S. Gonçalves. Smart Cities: Using Gamification and Emotion Detection to Improve Citizens Well Fair and Commitment. In Proceedings of the Intelligent Computing Conference (SAI 2020), pages 426–442, Virtual, 2020. Springer. [38] M. Rodrigues, V. Monteiro, B. Fernandes, F. Silva, C. Analide, and R. Santos. A Gamification Framework for Getting Residents Closer to Public Institutions. Journal of Ambient Intelligence and Humanized Computing, 11:4569–4581, 2020. [39] J. Ruohonen. Mysterious and Manipulative Black Boxes: A Qualitative Analysis of Perceptions on Recommender Systems. First Monday, 29(6):1–35, 2024. [40] J. Ruohonen. A Rapid Review Regarding the Concept of Legal Requirements in Requirements Engineering. Archived manuscript, available online: https: //doi.org/10.48550/arXiv.2509.06012, 2025. [41] J. Ruohonen, J. Koskinen, S. H. Klausen, and A. Gerdes. A Scenario Analysis of Ethical Issues in Dark Patterns and Their Research. Archived manuscript, available online: https://doi.org/10.48550/arXiv.2503.01828, 2025. [42] J. R. Saura, D. Palacios-Marqués, and A. Iturricha-Fernández. Ethical Design in Social Media: Assessing the Main Performance Measurements of User Online Behavior Modification. Journal of Business Research, 129:271–281, 2021. [43] C. Silva, C. Prandi, N. Nunes, and V. Nisi. Addressing Children and Caregivers’ Concerns in the Design of Child-Targeted Locative Systems. In Proceedings of the 21st International Conference on Mobile and Ubiquitous Multimedia (MUM 2022), pages 100–109, Lisbon, 2022. ACM. [44] I. Singh, G. Goyal, and A. Chandel. AlexNet Architecture Based Convolutional Neural Network for Toxic Comments Classification. Journal of King Saud University – Computer and Information Sciences, 34:7547–7558, 2022. [45] I. Stupavský and P. Dakić. Antisocial Behavior and the Dopamine Loop on Different Technological Platforms and Industries: An Overview. In Proceedings of Eighth International Congress on Information and Communication Technology (ICICT 2023), pages 471–481, London, 2023. Springer. [46] R. M. Tawafak, R. M. Tawafak, L. H. Al-Obaydi, and M. Pikhart. Competency Categorization and Roles of Online Teachers from the Perspective of University Students. Frontiers in Psychology, 14:1–12, 2023. [47] R. Watts. Truth and Politics: Thinking About Evidence-Based Policy in the Age Of Spin. Australian Journal of Public Administration, 73(1):34–46, 2014. [48] B. P. Woolf. Building Intelligent Interactive Tutors. Morgan Kaufmann Publishers, Burlington, 2009. [49] T. Yang, X. Shi, Q. Zeng, Y. Yang, C. Xu, and H. Liu. Optimization Methods in Fully Cooperative Scenarios: A Review of Multiagent Reinforcement Learning. Frontiers of Information Technology & Electronic Engineering, pages 479–509, 2025.