Online Reporter
Dr Evans Sangomba
Everything AI
AT first glance, the convenience is irresistible: prompt an AI, receive a polished essay, report or presentation in seconds.
Yet a pioneering MIT study of 54 participants exposes a hidden cost: when we outsource our writing to large language models, our brains disengage, critical thinking wanes, and genuine learning suffers.
As Zimbabwe accelerates its digital transformation, the time has come to ask whether this convenience is worth sacrificing the intellectual rigour that underpins our nation’s future.
The MIT researchers compared three groups over four months: one relying solely on their thinking (“Brain-only”), another using a standard search engine, and the third tapping ChatGPT for writing assistance.
Functional MRI scans revealed that the Brain-only group showed higher brain connectivity, particularly in regions associated with critical analysis and memory consolidation, compared to both the Search and LLM groups.
Moreover, participants who used ChatGPT reported lower satisfaction with their essays and displayed impaired ownership of their work. In short, the more they leaned on AI, the less they seemed to learn.
One alarming finding was the diminished inclination to critically evaluate AI-generated content.
When participants received polished paragraphs and “opinions” from the LLM, they rarely questioned the accuracy or bias of those outputs.
The frictionless convenience lulled them into passive acceptance.
This phenomenon mirrors the echo chamber effect we see on social media: instead of actively seeking diverse perspectives, users receive what the algorithm deems “top” or “relevant,” reinforcing existing beliefs.
In Zimbabwe, where media pluralism is still maturing, an AI-driven echo chamber risks narrowing our intellectual horizons even further.
Equally troubling was the impact on linguistic mastery. Compared to the Brain-only group, LLM users spent less time crafting their essays, exhibited weaker command of vocabulary and struggled to quote their work verbatim.
Writing is not merely transcription; it is a cognitive exercise that solidifies understanding, refines argumentation and instils confidence.
When AI takes on this labour, students miss out on the iterative self-editing process that builds lifelong writing skills. Over time, this could erode academic standards from primary schools to universities.
In Zimbabwe’s exam-centric education system, where performance in national assessments carries immense weight, the promise of AI assistance can be seductive.
A well-crafted ChatGPT essay might boost a learner’s Grade Seven literacy scores or produce an A-level English paper that dazzles examiners.
But such shortcuts yield hollow achievements. When knowledge gaps emerge in subsequent years, say, a law student who never grappled with case-law analysis or an engineer unfamiliar with technical report writing, the false confidence gained from AI assistance becomes a liability.
Beyond academic settings, professionals in business, law, journalism and public service also risk blunting their critical faculties.
A mid-career manager drafting policy briefs or a reporter crafting investigative stories may lean on AI to save time.
The MIT study warns that even experienced users can become over-reliant, accepting plausible-sounding but factually incorrect statements, a phenomenon dubbed “hallucination.”
In a parliamentary debate or court proceeding, basing arguments on AI fabrications could have severe repercussions for governance and justice.
Zimbabwe is also contending with deepening digital divides.
While urban centres enjoy better connectivity, rural areas still struggle with intermittent power and patchy internet.
If students or civil servants fall back on offline rote learning to preserve cognitive engagement, those with reliable AI access might pull further ahead academically, but at the expense of their critical thinking. Equitable development demands that AI adoption be accompanied by strategies that bolster, not erode, intellectual resilience across all regions.
What can Zimbabweans do to guard against AI-induced learning deficits?
First and foremost, educators must revise teaching and assessment methods. Assignments should emphasise process over outcome: students could submit research journals, outlines and rough drafts alongside final essays. Oral examinations, presentations and in-class writing exercises can verify genuine comprehension.
By making the writing journey transparent, teachers discourage unchecked AI use and reward authentic effort.
Schools and universities should integrate digital-literacy modules that spotlight AI’s limitations. Students need hands-on experience in fact-checking AI outputs, recognising bias and cross-verifying sources. Exercises might involve deliberately prompting ChatGPT with ambiguous or controversial questions, then analysing where and why the model errs.
Such critical engagements will sharpen evaluative skills and transform AI from a black-box oracle into a tool whose products are always subject to human scrutiny.
At the policy level, the Ministry of Primary and Secondary Education and the Zimbabwe Council for Higher Education should issue guidelines on AI usage in classrooms.
These could range from outright bans on generative-AI models for examination sittings to more nuanced frameworks allowing AI for preliminary research only, with mandatory human editing.
Clear policies help teachers enforce standards consistently and reassure parents that student learning remains paramount.
Parents also have a vital role.
Home-study routines must balance technology-assisted research with analogue exercises, reading physical books, handwriting essays and engaging in family debates.
By modelling deep reading and iterative writing practices, parents reinforce the message that true learning demands active mental participation. Discussions on the dinner table about current events, framed around news articles rather than AI summaries, help children cultivate independent viewpoints.
Zimbabwe’s teacher-training colleges can equip new educators with strategies to manage AI in the classroom. Workshops led by digital-pedagogy experts can showcase low-tech assessments, such as in-class annotations or timed hand-written essays, that ensure students internalise key concepts. Mentors can also encourage collaborative grading: teachers compare AI-augmented essays with student-only drafts to highlight the divergence in depth, originality and critical insight.
In higher education, research supervisors must insist on transparent methodology sections.
Theses and dissertations should specify whether generative AI was used for drafting, data analysis or literature review.
Supervisors can require students to archive their prompt histories and AI-output versions, enabling examiners to trace the evolution of the research argument. Such practices discourage over-reliance on AI and protect academic integrity.
Employers and professional bodies, ranging from the Law Society of Zimbabwe to the Institute of Engineers, should update codes of conduct to address AI usage.
Professionals who present AI-generated reports as their own work risk disciplinary action. Instead, organisations can offer internal training on how to leverage AI responsibly: using it for initial brainstorming, then critically refining outputs through peer review.
By embedding transparency, businesses signal that human expertise remains the ultimate authority.
On the regulatory front, the Ministry of Information Communication Technology and Courier Services, in collaboration with the Postal and Telecommunications Regulatory Authority (POTRAZ), should convene an AI advisory council. This body, comprising educators, technologists, ethicists and student representatives, can
develop national guidelines for responsible generative-AI use in education and industry. By anchoring policy decisions in multidisciplinary expertise, Zimbabwe ensures its frameworks are fit for local realities.
Public awareness campaigns are equally important. Just as we educate drivers on road safety, we must inform citizens about AI’s cognitive hazards. Radio programmes, community forums and social-media infographics can highlight the MIT study’s key takeaways, demonstrating how easy shortcuts today can stunt intellectual growth tomorrow. When families and communities grasp the stakes, they can collectively uphold norms that safeguard learning.
Critics may argue that resisting AI in writing is futile, that global trends cannot be reversed. Yet the goal is not to banish AI entirely but to harness it mindfully. Generative models can expedite idea generation, translate technical terms and suggest alternate phrasings. But the pen ultimately belongs to the human hand, or its digital equivalent. When students craft their sentences, wrestle with complex concepts and defend
arguments orally, they build the neural pathways essential for long-term mastery.
The MIT study serves as a wake-up call. Its findings echo concerns voiced by educators worldwide: that generative-AI tools, while powerful, can become cognitive crutches if left unchecked. In Zimbabwe’s
context, where education is both a personal lifeline and a cornerstone of national progress, the imperative is even stronger. We must cultivate a culture that values deep learning over quick fixes, critical inquiry over passive consumption and authentic creation over polished imitation.
In practical terms, this means adopting a balanced approach. Allow AI as an assistant, not a substitute. Use it to brainstorm essay topics, to proofread grammar or to explore unfamiliar vocabulary. But always require human oversight: confirm facts through trusted sources such as textbooks, academic journals and expert interviews. Encourage students to draft in their own words first, then compare and refine with AI suggestions.
When the next generation of Zimbabwean leaders graduate from schools in Masvingo to universities in Mutare, they must emerge with robust critical thinking, well-honed writing skills and a healthy scepticism toward convenient answers. Only then can they tackle complex challenges, from climate adaptation to equitable growth. In this endeavour, our embrace of AI must be tempered by a steadfast commitment to human intellect.
About the Author: Dr Evans Sagomba is a Doctor of Philosophy and Chartered Marketer (CMktr, FCIM) with an MPhil and PhD in Philosophy. He specialises in AI, Ethics, and Policy Research, and is an AI Governance and Policy Consultant. His expertise extends to Ethics of War and Peace, Philosophy of Development, and Political Philosophy. [email protected]. ORCID: 0009-0007-0681-0329.
Social media handles; LinkedIn; @Dr. Evans Sagomba (MSc Marketing)(FCIM )(MPhil) (PhD), X: @esagomba.
