Sciences Po's Policy on Artificial Intelligence

Our Policy on AI

Artificial intelligence (AI), particularly generative AI, already plays a significant role in both professional contexts and everyday life. Its rapid development is opening up a space of transformative change in our societies, our jobs, and our ways of life.

For Sciences Po, whose primary mission is to educate students to think critically about the contemporary world and act meaningfully within it, the challenge is twofold. We must ensure that our entire community – students, faculty, and staff – assimilates these tools so as to exploit their potential, while strengthening the requirement for discernment, creativity, depth of inquiry, critical thinking, and multidisciplinary understanding, which are essential skills for the judicious use of these tools.

The key concern is for our students to use AI as a lever to deploy the skills they have gained at Sciences Po, rather than as a substitute for their own judgement and creativity.

Strengthening cognitive skills and abilities

AI is changing the grammar of knowledge: what was expected of students until recently – analytical skills and the ability to synthesise information – is now seemingly within reach of a prompt.

The judicious use of AI therefore makes certain cognitive skills and abilities all the more fundamental:

• Discernment, to critically assess and exercise judgement over AI outputs, not only to avoid the pitfalls inherent in these tools, but also to utilise their full potential;
• Creativity, to move beyond the combinatorial logic of AI models;
• Intellectual ambition, to maintain the capacities for reading, writing, and concentration, the command of which is a prerequisite for thought itself.

These skills form the foundation of the intellectual education that Sciences Po provides, promotes and intends to strengthen.

A commitment to multidisciplinarity and the humanities

The multidisciplinary education delivered at Sciences Po brings different perspectives, knowledge, and methods into dialogue. This broad-based educational approach encompassing a plurality of disciplines and viewpoints is a major asset in handling AI. It allows students to develop the capacities for discernment, interpretation, contextualisation and critical distance that the proliferation of synthetic content renders essential.

Sciences Po also intends to strengthen its commitment to the humanities by deepening engagement with reading, writing, and artistic practices (particularly within the Institute for the Arts and Creation), which provide a foundation for reflection and stimulate the imagination and sensitivity that are intrinsic to human creativity.

Providing support and training for informed and responsible use of AI

With the workplace already being transformed, Sciences Po is actively engaged in the exploration and use of AI technologies for the benefit of its community.

Our approach is responsible and transparent. Sciences Po gives all members of its community the possibility to:

• test, compare and contrast different generative AI models, while ensuring that no sensitive data is shared;
• make judicious use of AI models while remaining mindful of their environmental cost and of the ethical use framework, including systematically acknowledging AI-generated content;
• encourage the use of open and transparent models whose limitations and biases are better understood;
• develop technical and critical literacy.

Training modules and resources are gradually being incorporated into our degree programmes at Bachelor’s, Master’s and Doctoral levels to promote the appropriate use of these tools, with due attention to the potential risks, biases and possibility of hallucinations inherent in these technologies.

Adapting assessment methods in the AI era

The use of AI in students’ academic work is widespread but not systematic. Sciences Po recognises this and has adopted a lucid approach: AI is neither to be banned nor ignored, but framed within clear guidelines while educating students on the imperatives of academic integrity.

This reflection, led by our teaching teams, has led to a gradual and differentiated evolution in assessment methods to adapt them to this new reality and make it possible to measure and recognise the skills acquired.

Assessment methods are defined according to the level of study and learning objectives of each programme and course. They may include traditional formats (written exams, oral exams, timed exercises) and formats that explicitly incorporate the use of AI (comparative analyses, AI-augmented assignments, reflection on biases). Students must then document their AI use by detailing the steps they took, specifying the tools they used, and adopting a reflective approach.

Whatever the chosen format, the assessment must measure students' ability to define problems, construct arguments, and contextualise; in other words, to produce independent thought.

Sciences Po encourages and supports teaching staff in testing new formats and sharing their practices, which are informed by their own professional experience and their understanding of employer expectations for our graduates. In this context, guidance on the use of generative AI for Sciences Po instructors has been published. Our Institute for Skills and Innovation, together with staff at the graduate schools, assists teachers in adapting assessment methods through training sessions, practical guides, and communities of practice.

What AI is doing to the world: research subjects at Sciences Po

With the breadth of its expertise and its ambitious research agenda, Sciences Po is establishing itself as a prime site for examining the rise of AI in all its dimensions and its impact on our societies: geopolitical, economic, societal, legal, and ethical issues, as well as implications for governance, regulation, and democratic processes. This effort is inseparable from the contribution of our researchers to the development of AI applications informed by knowledge specific to the humanities and social sciences.