{"id":5867,"date":"2026-02-25T14:00:39","date_gmt":"2026-02-25T17:00:39","guid":{"rendered":"https:\/\/blog.scielo.org\/en\/?p=5867"},"modified":"2026-02-25T14:01:55","modified_gmt":"2026-02-25T17:01:55","slug":"the-professors-dilemma-in-the-age-of-ai-do-we-teach-the-prompt-or-the-scientific-process","status":"publish","type":"post","link":"https:\/\/blog.scielo.org\/en\/2026\/02\/25\/the-professors-dilemma-in-the-age-of-ai-do-we-teach-the-prompt-or-the-scientific-process\/","title":{"rendered":"The professor’s dilemma in the age of AI: do we teach the prompt or the scientific process?"},"content":{"rendered":"

By Ricardo Limongi<\/strong><\/p>\n

\"Photograph<\/a>

Imagem: Berke Citak by Unsplash<\/a>.<\/em><\/p><\/div>\n

In a recent training session on artificial intelligence for researchers, a scene repeated itself. When explaining how a language model such as probabilistic logic that generates text works, and why the output needs to be verified, the audience dispersed. Minutes later, when demonstrating a prompt capable of reviewing academic paragraphs, everyone took notes. The same room, the same audience, two completely different levels of attention. This episode, far from being anecdotal, illustrates a dilemma that runs through the training of contemporary researchers: when we teach artificial intelligence (AI) in academia, are we teaching how to think with the tool or just how to operate it?<\/p>\n

The question is not trivial. The adoption speed of generative AI tools in scientific research has generated a legitimate demand for technical training. Researchers want and need to know how to use these technologies. The problem arises when training is reduced to teaching shortcuts, without understanding the underlying processes that give researchers the ability to critically evaluate what the tool produces.<\/p>\n

Instrumental literacy versus critical literacy<\/h3>\n

Recent literature on AI literacy presents a key distinction for understanding this dilemma. Long and Magerko (2020), in What is AI Literacy? Competencies and Design Considerations<\/a>1<\/sup><\/a><\/em>, in their seminal work presented at the CHI Conference, defined AI literacy as \u201ca set of competencies that enables individuals to critically evaluate AI technologies, communicate and collaborate effectively with AI, and use AI as a tool.\u201d This definition, widely adopted in subsequent literature, goes far beyond the operational scope of specific tools.<\/p>\n

Walter (2024), in Embracing the future of Artificial Intelligence in the classroom: the relevance of AI literacy, prompt engineering, and critical thinking in modern education<\/a>2<\/sup><\/a>,<\/em> a study published in the International Journal of Educational Technology in Higher Education, argues that the integration of AI in education requires not only technical skills but also critical thinking about the functioning and impacts of these technologies. The author proposes that AI literacy, prompt engineering, and critical thinking form an inseparable triad in contemporary education and that dissociating the first from the others compromises education.<\/p>\n

This distinction can be articulated on two levels. The first, which we can call instrumental literacy, corresponds to mastery of tools: knowing how to formulate prompts, understanding platforms, and performing tasks with the help of AI. It has practical value and responds to immediate demand. The second, critical literacy, corresponds to understanding what is at stake: understanding that a language model operates by statistical probability, not by semantic understanding; that \u201challucination\u201d is not an occasional defect, but a structural feature of systems that were not designed to produce truth, but rather statistically plausible text; that, in the scientific context, this distinction is not a technical detail but an epistemological foundation.<\/p>\n

Empirical evidence: understanding improves usage<\/h3>\n

Empirical research supports the idea that these two levels are not independent. Knoth et al. (2024), in AI literacy and its implications for prompt engineering strategies<\/a>3<\/sup><\/a>, <\/em>a study published in Computers and Education: Artificial Intelligence, demonstrated that students with greater conceptual AI literacy formulate better quality prompts and critically evaluate the results obtained. In other words, those who understand how the tool works use it more effectively than those who only master its commands.<\/p>\n

Brown, Sillence, and Branley-Bell (2025), in AcademAI: Investigating AI Usage, Attitudes, and Literacy in Higher Education and Research<\/a>4<\/sup><\/a>, <\/em>research published in the Journal of Educational Computing Research, investigated perceptions of AI among university students and faculty and identified that the lack of institutional guidance is one of the main barriers to responsible use. Participants highlighted the need for support to promote responsible use, showing that technical competence alone does not guarantee appropriate practices. The research also revealed that older age correlates with lower AI literacy, suggesting that training needs to reach not only students but also the trainers themselves.<\/p>\n

These findings converge with what Hackl et al. (2026), in The AI literacy heptagon: A structured approach to AI literacy in higher education<\/a>5<\/sup><\/a>, systematized when they identified 12 core competencies for AI literacy, organized from basic knowledge to practical skills, such as prompt engineering and ethical awareness. The proposed framework shows that prompt engineering is one of the competencies, not the core competency, nor certainly the only one.<\/p>\n

The dilemma in the classroom<\/h3>\n

A recent episode illustrates the practical dimension of this problem. An experienced professor and active advisor reported asking ChatGPT to recommend a scientific article for his advisee. He received the title, authors, journal, and year, all formally impeccable. The article, however, did not exist. The question that followed was straightforward: \u201cIf AI hallucinates, when can I trust it?\u201d<\/p>\n

The technical answer is simple: at no time, without independent verification. But the episode reveals something deeper. A researcher with decades of experience treated a generative model as a factual database. Not out of negligence or general ignorance, but out of a lack of specific literacy about what that tool is and, more importantly, what it is not.<\/p>\n

This is where the professor’s dilemma becomes more acute. Educators who recognize the need to teach critical literacy face structural resistance: the contemporary information ecosystem(social media, quick courses, viral content) systematically reinforces the logic of shortcuts. When the content is conceptual, engagement disperses; when it becomes practical advice, attention is focused. Professors, pressured to demonstrate relevance and generate immediate results, feel the legitimate temptation to give in to demand and teach only the prompt<\/em>.<\/p>\n

New researchers: interest or indifference?<\/h3>\n

The most uncomfortable question underlying this dilemma is whether new researchers will be interested in conceptual training in AI. The honest answer is that, in the short term, many probably will not. Contemporary academic culture, driven by productivity metrics, pressured by deadlines, and immersed in an ecosystem that rewards quick solutions, creates incentives that work against deepening knowledge.<\/p>\n

However, the history of methodological training offers an instructive parallel. Research methodology and statistics have rarely been among the most popular undergraduate subjects. And yet, researchers with solid methodological training distinguished themselves throughout their careers not because they mastered the execution of a specific statistical test, but because they understood what the test meant and when it was (or was not) applicable.<\/p>\n

AI literacy may follow a similar trajectory. Specific tools and prompts are, by nature, perishable: the speed of model evolution makes any specific command potentially obsolete in a matter of months. The ability to understand what happens between the question and the answer, that is, critical literacy, does not perish. Those who develop this understanding today will be prepared for tomorrow’s tools, including those that will dispense with prompts as we know them.<\/p>\n

From prompt engineering to literacy as institutional policy<\/h3>\n

In Navigating the landscape of AI literacy education: insights from a decade of research (2014\u20132024)<\/a>6<\/sup><\/a>, <\/em>\u00a0an integrative review, conducted by researchers and published in Humanities and Social Sciences Communications, mapped the evolution of the field of AI literacy over the last decade (2014\u20132024) and identified a persistent gap: despite the exponential growth of publications on AI in education, conceptual training remains secondary to technical-instrumental training \u00a0The field has evolved in volume, but not necessarily in depth.<\/p>\n

This finding has direct implications for educational and research institutions. If AI literacy is, as the literature suggests, a condition for the effective and responsible use of these tools, then its promotion cannot depend solely on individual initiatives by sensitized professors. It needs to be treated as an institutional issue integrated into curricula, graduate programs, and professor training policies.<\/p>\n

In Brazil, this discussion takes on specific contours. As Diretrizes para o uso \u00e9tico e respons\u00e1vel da Intelig\u00eancia Artificial Generativa<\/a>7<\/sup><\/a>, published by Sampaio, Sabbatini, and Limongi (2024), already propose essential competencies, such as understanding the tools and their limitations, maintaining human authorship as a central element, and critically evaluating outputs. SciELO, through its Guide to the Use of AI Tools and Resources<\/a>8<\/sup>, has established clear principles for declaration and verification. These initiatives, however, need to be translated into concrete training practices.<\/p>\n

Concluding remarks<\/h3>\n

The dilemma facing professors in the age of AI cannot be resolved by choosing between teaching prompts and teaching process. Prompts are tools, and tools matter. The question is whether we train researchers who understand what they are doing or operators who know where to press buttons.<\/p>\n

The available evidence suggests that critical literacy does not compete with instrumental literacy; it enhances it. Researchers who understand how tools work use them better, evaluate their results more rigorously, and adapt more easily to technological changes. Focusing exclusively on shortcuts creates professionals who are dependent on specific tools; focusing on literacy trains researchers who are able to navigate a constantly changing landscape.<\/p>\n

Notes<\/h3>\n

1. LONG, D. and MAGERKO, B. What is AI Literacy? Competencies and Design Considerations.Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems<\/em> [online]. 2020, vol. 2020, no 191, pp. 1\u201316 [viewed 27 February 2026]. https:\/\/doi.org\/10.1145\/3313831.3376727<\/a>. Available from: https:\/\/dl.acm.org\/doi\/10.1145\/3313831.3376727<\/a>↩<\/a><\/p>\n

2. WALTER, Y. Embracing the future of Artificial Intelligence in the classroom: the relevance of AI literacy, prompt engineering, and critical thinking in modern education. International Journal of Educational Technology in Higher Education<\/em> [online]. 2024, vol. 21, no. 1, art. 15, ISSN: 2365-9440 [viewed 27 February 2026]. https:\/\/doi.org\/10.1186\/s41239-024-00448-3<\/a>. Available from: https:\/\/link.springer.com\/article\/10.1186\/s41239-024-00448-3<\/a>↩<\/a><\/p>\n

3. KNOTH, N.,et al.<\/em> AI literacy and its implications for prompt engineering strategies.Computers and Education: Artificial Intelligence<\/em> [online] 2024, vol. 6, art. 100225. [viewed 27 February 2026] https:\/\/doi.org\/10.1016\/j.caeai.2024.100225<\/a>. Available from: https:\/\/www.sciencedirect.com\/science\/article\/pii\/S2666920X24000262?via%3Dihub<\/a>↩<\/a><\/p>\n

4. BROWN, R., SILLENCE, E. and BRANLEY-BELL, D. AcademAI: Investigating AI Usage, Attitudes, and Literacy in Higher Education and Research. Journal of Educational Computing Research<\/em> [online]. 2025, vol. 54, no.1, ISSN: 0735-6331 [viewed 27 February 2026]. https:\/\/doi.org\/10.1177\/00472395251347304<\/a>. Available from: https:\/\/journals.sagepub.com\/doi\/10.1177\/00472395251347304↩<\/a><\/p>\n

5. HACKL, V., MULLER, A. E. and SAILER, M.. The AI literacy heptagon: A structured approach to AI literacy in higher education.Computers and Education: Artificial Intelligence <\/em>[online]. 2026, vol. 10, ISSN: 2666-920X. [viewed 27 February 2026]. https:\/\/doi.org\/10.1016\/j.caeai.2026.100540<\/a>. Available from: https:\/\/www.sciencedirect.com\/science\/article\/pii\/S2666920X26000019?via%3Dihub<\/a>↩<\/a><\/p>\n

6. YANG, Y., et al<\/em>.. Navigating the landscape of AI literacy education: insights from a decade of research (2014\u20132024). Humanities and Social Sciences Communications<\/em> [online]. 2025, vol. 12, no. 1, pp. 1-12. [viewed 27 February 2026]. https:\/\/doi.org\/10.1057\/s41599-025-04583-8<\/a>. Available from: https:\/\/www.nature.com\/articles\/s41599-025-04583-8<\/a>↩<\/a><\/p>\n

7. SAMPAIO, R.C., SABBATINI, M. and LIMONGI, R. Diretrizes para o uso \u00e9tico e respons\u00e1vel da Intelig\u00eancia Artificial Generativa: um guia pr\u00e1tico para pesquisadores. S\u00e3o Paulo: Editora Intercom, 2024. Available from: https:\/\/www.portcom.intercom.org.br\/ebooks\/detalheEbook.php?id=57203<\/a>↩<\/a><\/p>\n

8. Guia de uso de ferramentas e recursos de Intelig\u00eancia Artificial na comunica\u00e7\u00e3o de pesquisas na Rede SciELO [online]. SciELO \u2013 Scientific Electronic Library Online, 2023 [viewed 27 February 2026]. Available from: https:\/\/wp.scielo.org\/wp-content\/uploads\/Guia-de-uso-de-ferramentas-e-recursos-de-IA-20230914.pdf<\/a>↩<\/a><\/p>\n

References<\/h3>\n

BROWN, R., SILLENCE, E. and BRANLEY-BELL, D. AcademAI: Investigating AI Usage, Attitudes, and Literacy in Higher Education and Research. Journal of Educational Computing Research<\/em> [online]. 2025, vol. 54, no.1, ISSN: 0735-6331 [viewed 27 February 2026]. https:\/\/doi.org\/10.1177\/00472395251347304<\/a>. Available from: https:\/\/journals.sagepub.com\/doi\/10.1177\/00472395251347304<\/a><\/p>\n

Guia de uso de ferramentas e recursos de Intelig\u00eancia Artificial na comunica\u00e7\u00e3o de pesquisas na Rede SciELO [online]. SciELO \u2013 Scientific Electronic Library Online, 2023 [viewed 27 February 2026]. Available from: https:\/\/wp.scielo.org\/wp-content\/uploads\/Guia-de-uso-de-ferramentas-e-recursos-de-IA-20230914.pdf<\/a><\/p>\n

HACKL, V., MULLER, A. E. and SAILER, M.. The AI literacy heptagon: A structured approach to AI literacy in higher education.Computers and Education: Artificial Intelligence <\/em>[online]. 2026, vol. 10, ISSN: 2666-920X. [viewed 27 February 2026]. https:\/\/doi.org\/10.1016\/j.caeai.2026.100540<\/a>. Available from: https:\/\/www.sciencedirect.com\/science\/article\/pii\/S2666920X26000019?via%3Dihub<\/a><\/p>\n

KNOTH, N.,et al.<\/em> AI literacy and its implications for prompt engineering strategies.Computers and Education: Artificial Intelligence<\/em> [online] 2024, vol. 6, art. 100225. [viewed 27 February 2026] https:\/\/doi.org\/10.1016\/j.caeai.2024.100225<\/a>. Available from: https:\/\/www.sciencedirect.com\/science\/article\/pii\/S2666920X24000262?via%3Dihub<\/a><\/p>\n

LONG, D. and MAGERKO, B. What is AI Literacy? Competencies and Design Considerations.Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems<\/em> [online]. 2020, vol. 2020, no 191, pp. 1\u201316 [viewed 27 February 2026]. https:\/\/doi.org\/10.1145\/3313831.3376727<\/a>. Available from: https:\/\/dl.acm.org\/doi\/10.1145\/3313831.3376727<\/a><\/p>\n

SAMPAIO, R.C., SABBATINI, M. and LIMONGI, R. Diretrizes para o uso \u00e9tico e respons\u00e1vel da Intelig\u00eancia Artificial Generativa: um guia pr\u00e1tico para pesquisadores. S\u00e3o Paulo: Editora Intercom, 2024. Available from: https:\/\/www.portcom.intercom.org.br\/ebooks\/detalheEbook.php?id=57203<\/a><\/p>\n

WALTER, Y. Embracing the future of Artificial Intelligence in the classroom: the relevance of AI literacy, prompt engineering, and critical thinking in modern education. International Journal of Educational Technology in Higher Education<\/em> [online]. 2024, vol. 21, no. 1, art. 15, ISSN: 2365-9440 [viewed 27 February 2026]. https:\/\/doi.org\/10.1186\/s41239-024-00448-3<\/a>. Available from: https:\/\/link.springer.com\/article\/10.1186\/s41239-024-00448-3<\/a><\/p>\n

YANG, Y., et al<\/em>.. Navigating the landscape of AI literacy education: insights from a decade of research (2014\u20132024). Humanities and Social Sciences Communications<\/em> [online]. 2025, vol. 12, no. 1, pp. 1-12. [viewed 27 February 2026]. https:\/\/doi.org\/10.1057\/s41599-025-04583-8<\/a>. Available from: https:\/\/www.nature.com\/articles\/s41599-025-04583-8<\/a><\/p>\n

 <\/p>\n

About Ricardo Limongi Fran\u00e7a Coelho<\/h3>\n

\"Photograph<\/a><\/p>\n

Professor of Marketing and Artificial Intelligence, Universidade Federal de Goi\u00e1s (UFG), Goi\u00e2nia\u2013GO,and Editor-in-Chief of Brazilian Administration Review (BAR) da ANPAD journal, DT-CNPq Scholarship Recipient.<\/p>\n

 <\/p>\n

 <\/p>\n

 <\/p>\n

Translated from the original in\u00a0Portuguese<\/a>\u00a0by Lilian Nassi-Cal\u00f2.<\/p>\n","protected":false},"excerpt":{"rendered":"

The question is not trivial. The adoption speed of generative AI tools in scientific research has generated a legitimate demand for technical training. Researchers want and need to know how to use these technologies. The problem arises when training is reduced to teaching shortcuts, without understanding the underlying processes that give researchers the ability to critically evaluate what the tool produces. …<\/span> Read More →<\/span><\/a><\/span><\/p>\n","protected":false},"author":143,"featured_media":5871,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_links_to":"","_links_to_target":""},"categories":[3],"tags":[84,86,9],"class_list":["post-5867","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-analysis","tag-artificial-intelligence","tag-ethics-in-scientific-communication","tag-scielo-program"],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/posts\/5867","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/users\/143"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/comments?post=5867"}],"version-history":[{"count":6,"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/posts\/5867\/revisions"}],"predecessor-version":[{"id":5873,"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/posts\/5867\/revisions\/5873"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/media\/5871"}],"wp:attachment":[{"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/media?parent=5867"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/categories?post=5867"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.scielo.org\/en\/wp-json\/wp\/v2\/tags?post=5867"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}