Why Responsible AI in Lebanon Cannot Simply be Imported
Artificial intelligence did not arrive in Lebanon at a moment when institutions had the luxury of calmly asking what it might change. It entered while schools, universities, cultural organisations, businesses and families were already dealing with economic collapse, political paralysis, repeated crises and war, and perhaps that partly explains why its use has spread so quickly. At the same time, serious discussion about responsibility, governance and ethics remains, in everyday practice at least, remarkably thin.
Students use it for assignments, teachers for preparation, artists for images and texts, researchers for summaries, businesses for communication, journalists for drafting and people for everything from translation to advice, often with very little reflection on what is being shared with these systems, where the information comes from, whether the answer is reliable, what happens to authorship, or whether the apparent efficiency is quietly replacing skills we still need to preserve.
This gap has become increasingly central to my work. My own path into AI did not come through engineering, but through education, research, arts and culture, curriculum development, intercultural dialogue and consultancy, fields in which questions of representation, interpretation, power, responsibility and human agency have always mattered. Over the last few years, I have increasingly been working on AI ethics, bias, cultural accuracy and human-AI communication, particularly in educational and cultural contexts. What interests me therefore is not how well people can use AI, but whether they understand what they are handing over to it.
Last Saturday at Kulturnest, I worked with students on using AI in writing and the creative process, and the session did not begin with tricks for producing better prompts, because that is probably the easiest part to learn. Instead, we looked at what happens when a machine can produce a polished paragraph, propose a plot or complete an idea before a young person has had enough time to discover what they themselves wanted to write.
We discussed ways AI can be useful without becoming the author: helping to reorganise a draft, suggesting another angle, questioning an idea, proposing alternatives or supporting revision, while leaving the imaginative and intellectual decisions with the student, because there is a considerable difference between asking a machine to write a story for you and using it to challenge a story that has already begun to exist in your own mind.
That distinction matters, especially for younger learners, because writing is beyond the production of a clean final text. It's about the hesitations, bad sentences, discarded ideas, and unexpected turns that are part of how thinking develops. If every difficult stage is removed in the name of efficiency, we may eventually produce students who are very good at obtaining finished answers but increasingly unsure of what they can formulate without assistance.
The same concern has appeared in other sessions I have been developing at Kulturnest, although the emphasis changes with each audience: with university students, I have worked on AI ethics, academic writing, authorship, research integrity and fabricated references, while with creatives in Lebanon and abroad the conversation has turned more toward voice, originality, cultural context and the point at which assistance begins to feel like substitution.
This is one of the reasons I have become wary of speaking about “AI literacy” as though it were a single transferable package. A thirteen-year-old, a graduate student, a journalist, an artist and a museum curator may be using similar technologies, but the risks, responsibilities and ethical questions are not the same, and neither should be the training. And the same applies to governance.
International frameworks are necessary, but I do not believe that responsible AI practice in Lebanon can consist of importing a policy written somewhere else, translating it into Arabic, changing the institution’s name and assuming the work is finished, because our context brings its own difficulties: several languages used interchangeably, unequal access to resources, weak institutional structures, contested histories, fragile archives, political and religious sensitivities, and a cultural sector already functioning under enormous pressure.
A Lebanese student asking an AI system about the Civil War, Palestine, sectarianism, gender or religion is not retrieving “information”. The response is shaped by what was digitised, what was available in the training material, which languages were better represented, which terminology was dominant and which narratives were absent or marginal, and these absences are particularly important in a region where a great deal of knowledge still exists outside large searchable digital repositories.
For arts and culture, the issue becomes even more complex. After all, a painting, an archive, a religious image, a testimony or an artist’s visual language cannot be treated as neutral data because a machine is capable of absorbing it. A museum, journal, gallery or creative enterprise therefore needs more than a general statement about “responsible AI”. It needs to know what its teams are already doing with these tools, which material may be uploaded, what requires consent, how outputs are verified, when AI use should be disclosed, how copyright and confidentiality are handled, which decisions remain human, and who is accountable when something goes wrong.
Lebanon makes this particularly urgent because improvisation is almost a way of life here, and while that ability to work around broken systems often produces remarkable creativity, it also means that new technologies can become normal long before anybody has decided what limits should accompany them. And war intensifies this tendency. When institutions are understaffed, people are exhausted and time and money are scarce, anything that promises speed becomes attractive, which is precisely why questions about verification, privacy, authorship and responsibility can so easily be postponed; yet waiting for a calmer Lebanon before discussing responsible AI may mean waiting indefinitely, while habits are already being formed every day in classrooms, offices, studios and homes.
I am not interested in teaching fear of AI, nor in celebrating every new capability simply because it exists. What matters to me is helping people understand when the technology is genuinely extending their capacities and when, almost unnoticed, it is beginning to replace judgment, learning, voice or responsibility. In that perspective, the conversation cannot be imported ready-made, but it has to be developed here, with the people who will actually live with its consequences.
Dr. Pamela Chrabieh, Kulturnest co-founder & CEO
---------------------------
CLICK HERE for further information about our AI clinics for students, academics and creatives.