Equipping Psychology Majors with AI Proficiency Will Empower Them,but Will It Single Them Out? Comment on Robinson et al. (2026)

In their target article, Robinson et al. (2026) propose that integrating resources provided by the American Psychological Association (APA) with artificial intelligence (AI) proficiency will empower the psychology major, thereby boosting successful graduates’ chances on the job market. Not only will it increase their chances for traditional jobs but it will also open new opportunities for them beyond typical psychology occupations. Robinson et al. refer to different resources and illustrate how to embed AI proficiency in each of them. They conclude that existing psychology major programs need to redesign their curricula, thereby enabling students to enhance their cognitive, communication, personal, social, and technological skills by integrating AI.

Without a doubt, AI is having immense effects on the current and future labor market and the skill sets required. According to job market analyses, it will be replacing humans in many routine and administrative tasks, and—if used proficiently—can be used to accelerate human performance in complex tasks that require reasoning, deliberation, or problem-solving. In the latter context, AI systems can be considered “partners in cognition” (Salomon et al., 1991), in which labor is divided among the human and artificial partners, whose interaction develops over time and becomes increasingly interdependent. According to Salomon et al., as humans collaborating with an AI system, we need to be aware of the capabilities and fallacies of our digital partner and take full responsibility for the appropriateness of the outcomes, even if they were jointly achieved. Thus, to benefit from AI, first and foremost, we need to engage in what Salomon et al. describe as mindful processing. “For partnering to attain higher levels of intellectual performance, the human's mental processes have to be of the nonautomatic type … the employment of such nonautomatic, effortful, and thus metacognitively guided processes has been defined as a state of mindfulness.” Only when a human functions mindfully “will the upgrading of performance while working with a computer tool take place” (Salomon et al., 1991, p. 9). This quote is more than 30 years old. Nevertheless, to me, it signifies the unique position of psychological science to advance our understanding of what it requires to benefit from using AI (even though Gavriel Salomon was an educational scientist by training).

When reading the contribution of Robinson et al. (2026), I was missing a discussion of this unique relationship between AI and psychology. I fully agree that psychology majors should know how to use AI proficiently and in ethical ways. But so should anybody obtaining an academic degree. Hence, AI proficiency training needs to be incorporated in every study program to prepare students for a job market that will become increasingly influenced by AI. This is because the skills described as part of The Skillful Psychology Student profile (Naufel et al., 2018) are relevant to almost every profession, albeit to varying degrees. However, that calls into question whether simply by integrating AI proficiency into psychology curricula, respective majors will have better chances on the job market than their competitors from other (related) disciplines. My concern is that this will rather end in a rat race between psychology and other disciplines, with some being faster in integrating AI proficiency into their programs than others.

Admittedly, Robinson et al. (2026) present some arguments for why psychology might win such a race, thereby putting our discipline into a pole position. In particular, they argue that psychology can serve as a role model because of its profound understanding of what constitutes learning and good teaching, respectively. In principle, I agree that our discipline is in a good position in this regard—but I do not yet see this knowledge being fully used in practice. First, in my view, other disciplines (and not just the obvious candidates such as computer science or computational linguistics) have actually been much faster to incorporate AI proficiency into their curricula than psychology—including medicine, economy, STEM, or disciplines involved in teacher education. This is because AI entered their fields of occupation at such a speed that an adjustment of higher education programs was unavoidable. For instance, in schools, the majority of pupils use AI both in productive and unproductive ways, causing a severe disruption to existing teaching and assessment routines, so that teacher education and professional development programs had to respond quickly to prepare (preservice) teachers to consider AI in their teaching, both as a tool and as a subject. In contrast, in my experience, psychology programs have been rather slow in responding to the ubiquitous availability of AI. Second, I am concerned that even though psychologists should know what constitutes good learning and teaching, this is not necessarily reflected in existing curricula and instruction offered in psychology. Even the present article emphasizes that students should be enabled to show efficient task performance using AI, for instance, when analyzing data or creating automated research protocols and results documentation, to then be able to follow career paths as assistants in various fields of occupation (see Table 3 in Robinson et al., 2026). To me, this does not resonate well. I want future psychologists to develop solutions for nonroutine problems as well as deal with scenarios that require deep reasoning, thinking, and complex decision-making based on a profound and multidimensional understanding of human behavior. I fear that many of the cognitive skills described for the Baccalaureate level in Table 3 will become subject to automation and, at a nearby stage of AI development, be carried out completely or to a large degree by AI systems, making the kind of psychologists envisioned by Robinson obsolete rather than assets on the job market for which there is high demand. As a consequence, we need to teach our students to become deep thinkers and problem-solvers rather than efficient task-performers.

Therefore, I suggest a complementary approach that will help to increase the perceived aptitudes of psychology majors for the job market and, at the same time directly support APA's mission of also enhancing the public impact of psychology in society (APA, 2023). Rather than only teaching students to become proficient AI users (which we should definitely do), we should strengthen their competences in using the wealth of theories, models, empirical results, and methods of psychological science to contribute to a better understanding of when and how to use AI, with the mission of developing a psychology-driven, human-centered perspective on AI. This training strategy should be accompanied by increased public engagement of psychological scientists in debates concerning the potentials and risks of AI for society, with a clear message that if the psychology of users is ignored, the ubiquitous use of AI is likely to cause more harm than benefits.

Notably, there are so many examples where the introduction of technological innovations has failed even when it was guided by the best of all intentions to improve the well-being of mankind, because psychological factors were not taken into account sufficiently. Assembly-line work, the way it was initially envisioned by Henry Ford, failed because workers perceived the work as overly monotonous and committed errors. Medical treatments often do not have the intended positive effects, because patients do not comply with the instructions given, forget to take their medication, or do not continue exercising once they feel better. Educational environments that are aimed at promoting learners’ autonomy and self-efficacy beliefs often overwhelm students because they lack the necessary cognitive strategies and cannot monitor and regulate their learning to the extent necessary in these environments. Many of these problems could have been avoided if the users’ psychological processes (i.e., their motivation and volition, emotion, cognition, and social processes) had been considered.

Thus, in my view, the unique selling point of psychological science that we should aim to further develop in our students is the ability to describe, explain, predict, and modify human behavior when humans are confronted with technological innovations such as AI.

Let me illustrate this claim for the case of using AI in education, where we already know quite a bit about human behavior when learners interact with AI systems. For instance, many of the AI-based learning scenarios that have been deployed so far serve human's tendencies for cognitive offloading (Risko & Gilbert, 2016). Thus, students no longer invest the mental effort necessary to develop skills and knowledge; rather, they delegate cognitive tasks to the AI. According to Bauer et al. (2025), these inversion effects where cognitive engagement is reduced rather than promoted, inhibit learning from AI and can lead to skill skipping (not developing new skills) or de-skilling (unlearning or not consolidating learned skills). Moreover, because of the probabilistic and potentially biased nature of AI-generated output, students need to reflect upon its correctness, adequacy, and helpfulness when learning with an AI-based system; however, students have been shown not to engage in these metacognitive processes. That is, they overly rely on and trust AI-generated output, a phenomenon called metacognitive laziness (Fan et al., 2024). Also, AI systems tend to confirm users’ opinions and beliefs even when users err, express unethical intentions, or propose to do harm to themselves or others (Cheng et al., 2026). This sycophancy can have very negative effects regarding the users’ real-life behavior (e.g., increased likelihood of committing illegal actions or suicide). In educational contexts, AI sycophancy is likely to increase perceived fluency concerning the interaction with the system, which in turn may promote the learners’ overconfidence regarding their knowledge (Bjork et al., 2013) and reduce their future engagement in learning. As a consequence, AI systems should be used in education so that cognitive offloading is avoided, metacognitive reflection is promoted, and high-quality feedback is guaranteed. In this vein, Bauer et al. (2025) present the Inversion-Substitution-Augmentation-Redefinition model, according to which the use of AI in educational contexts will have different effects on learning outcomes, depending on whether it suppresses or displaces helpful learning processes (Inversion), replaces them (Substitution), makes them more effective (Augmentation) or promotes them in a way that would be impossible without AI (Redefinition). It shows how a profound understanding regarding the psychology of learning—as one aspect of the curriculum for psychology majors—may help to leverage the potential of AI in the application context of education. Similar insights exist in all the various subdisciplines of psychology that could be brought to action to better understand and to design effective ways of using AI at work, in medical care, or for leisure.

In line with this observation, the German Psychological Association launched an initiative in 2025 to draw on the broad psychological expertise of its members that will help to better understand the impact of AI on society and the individual and allow for evidence-based recommendations for its responsible use to public discourse (https://www.dgps.de/schwerpunkte/arbeitsgruppe-ki-und-psychologie/; available in German only). The guiding principle of the established working group is that only by taking into account psychological insights from all subfields can AI systems be designed to offer added value for individuals, groups, and society, take diverse human needs and perspectives into account, and at the same time minimize risks as much as possible. The working group brings together experts from various subdisciplines of psychology, including industrial and organizational psychology, engineering psychology, clinical psychology, media psychology, and educational psychology, to ground its recommendations in a comprehensive view of psychological science. Its intention is to use respective insights not only to foster understanding but also to shape technological innovations responsibly by supporting researchers, developers, and policymakers in creating AI systems that are scientifically sound and valuable to society. To achieve its aims, the initiative prepares policy papers and organizes a conference that will bring together psychological scientists and stakeholders from politics, education, and occupational fields to discuss ways of using AI in a human-centered manner. Moreover, there will be a Special Issue published in 2026 in the society's central outlet (Psychologische Rundschau: https://www.hogrefe.com/de/zeitschrift/psychologische-rundschau), which will include short position papers from various subdisciplines in psychology to discuss their perspectives on AI.

To conclude, I am convinced that psychology majors have very good chances on the job market. But this is not only because we as teachers add AI proficiency to our competence profiles. Rather, we need to highlight in our teaching as well as when engaging in public debates that psychological science is an indispensable ingredient for the development and implementation of any AI system.