The rise of artificial intelligence in higher education compels us to rethink our writing practices and assessment criteria. This text, developed as part of a reflection at UNIL and presented during Compilatio’s AI Week, traces the history of ‘writing companions’ and introduces a map of writing competencies, enriched by the AIAS scale. It is intended for educators, students, and academic leaders who wish to uphold academic integrity while thoughtfully integrating AI. An invitation to open a pedagogical dialogue on what it means to produce ‘authentic’ writing today.
The emergence of artificial intelligence in academia raises fundamental questions about our writing and assessment practices. As part of our ongoing reflections on the impact of AI in education, we have humbly begun an inquiry that takes both a historical and forward-looking perspective.
This analysis, initially presented during Compilatio’s AI Week (2024), offers a preliminary exploration of the topic through a central question: how can we conceptualize the authenticity of academic work in an environment where AI is becoming a new ‘writing companion’?
Our approach is based on three main pillars:
- A historical perspective on writing practices, demonstrating that the act of writing has never truly been a solitary endeavor.
- A mapping of academic writing skills, based on the Science Writing Toolbox (2024) developed at UNIL by Prof. Preitner.
- A methodological proposal outline using the AIAS scale for the rational integration of AI into teaching practices
This reflection is intended for academics, educational coordinators, and students seeking to navigate the evolving landscape of academic practices. It aims to open avenues for preserving academic integrity while embracing the opportunities offered by emerging technologies.
Far from adopting a technophobic or technophilic stance, our inquiry seeks to develop a nuanced approach grounded in the realities of contemporary pedagogical practices. It highlights the need to rethink our assessment criteria in order to value distinctly human skills, while acknowledging the transformative potential of AI in higher education.
Historical perspective of writing companions
The evolution of our writing practices reflects a constant adaptation to technological and social innovations. Oral tradition, the earliest form of knowledge transmission, relied on direct and collective experience. The advent of manuscripts introduced the first form of mediation, with scribes serving as privileged custodians of knowledge. The printing revolution marked a decisive turning point, democratizing access to knowledge and standardizing its dissemination. The industrial era multiplied sources of information, leading to an unprecedented diversification of available references. The digital age then radically transformed our relationship to information, offering near-unlimited access to knowledge. Today, AI continues this historical trajectory, emerging as a new ‘writing companion’ that, like its predecessors, raises fundamental questions about authenticity and intellectual production.
Mapping academic writing skills
This competency map, inspired by the Science Writing Toolbox (2024), provides a rigorous analytical framework for understanding the multifaceted nature of academic writing. Each identified skill represents a distinct yet interconnected aspect of the writing process. Planning and structuring form the methodological foundation, enabling coherent organization of thought. Critical analysis and synthesis reflect the ability to process information meaningfully. Writing and audience adaptation underscore the importance of targeted communication. Narrative creativity enriches academic discourse with an engaging dimension. Ethics and methodology ensure scientific rigor, while revision supports continuous improvement. This systemic approach helps pinpoint potential areas for AI intervention, while safeguarding the integrity of the academic process.
Distinction between AI capabilities and distinctive human skills
Differentiating between AI capabilities and distinctly human competencies is essential for rethinking our assessment practices. AI models excel at generating structured content and analyzing patterns, demonstrating notable efficiency in systematic and repetitive tasks. However, human skills stand out through their unique ability to contextualize, innovate, and forge novel connections. This dichotomy should not be seen as a conflict, but rather as an opportunity to redefine our evaluation criteria. By clearly identifying what can be automated and what remains within the realm of human expertise, we can adapt our pedagogical methods to better value the truly distinctive competencies of our students.
The AIAS scale as a tool for dialogue
- Level 0 : No AI
- Level 1: Structuring and brainstorming
- Level 2: Human production enhanced by AI
- Level 3: Human-enhanced AI production
- Level 4: Full AI integration
The AIAS scale, developed by Perkins, Fuze, Roe, and Mac Vaught (2024), offers a valuable methodological framework for fostering constructive dialogue between instructors and students regarding the use of AI. This graduated model enables transparent communication of expectations and permissions concerning the use of AI tools in academic work. Each level corresponds to a specific degree of integration, providing flexibility aligned with diverse pedagogical goals. This structured approach not only clarifies expectations but also supports fairer and more consistent evaluation of student work, while maintaining the necessary academic rigor.
Educational implications and practical recommendations
Questioning our pedagogical practices in the face of AI’s emergence requires a methodical and thoughtful approach. The integration of these tools into higher education cannot occur without structured support for all stakeholders. Training in the meaningful use of AI must go hand in hand with a deeper reflection on our assessment methods. The challenge is not merely technical, but fundamentally pedagogical: how can we foster and value distinctly human competencies in an environment where AI is becoming a daily tool? The answer lies in a balanced approach—one that combines technological innovation with the preservation of the core values of higher education.
Conclusion: Revisiting the concept of academic authenticity in the age of AI
The emergence of AI as a writing companion marks a new stage in the evolution of academic practices. Rather than a rupture, this transformation continues the long-standing tradition of higher education adapting to technological innovation. We believe the core challenge lies not in resisting or uncritically embracing these tools, but in our ability to rethink pedagogical and assessment approaches.
Professors and researchers are now called upon to develop new skills—not only technical, but above all pedagogical and methodological. The AIAS scale and the proposed competency map serve as structuring tools to support this transformation. A clear distinction between AI capabilities and distinctly human competencies allows us to focus our efforts on valuing what remains fundamentally human: creativity, critical thinking, and the ability to contextualize.
This reflection must be enriched by the contributions of true specialists in the field, and we will seek input from expert colleagues at UNIL to better frame the discussion and offer our academic community more valid and in-depth perspectives on the challenges surrounding scientific writing in the age of AI.