EDITORIAL
STRATEGY
Trusted content, built for AI search.
BECOME THE SOURCE.
AI has opened a gap between what's easy to produce and what's worth trusting. And it is changing what gets found and how that gets told.
I work in that gap. It has terrible lighting.
My background crosses business, political and social sciences, documentary arts and multilingual content creation. What connects all of it is the same instinct: making expertise and difficult knowledge legible and compelling in whatever form serves the ideas best.
Today that work means staying close to what AI is doing to research, writing, creativity, discovery, audience expectations and credibility. For organisations operating in the European Union, specifically, that now includes rules on marking and disclosure.
EXAVOICE grows out of that intersection: editorial judgment informed by ongoing inquiry into how AI is reconfiguring the information ecosystem.
AI horizon tracking
-
Answer engines now sit between an organisation’s expertise and its audience. I track how they select, summarise, cite, and represent (or misrepresent) knowledge and turn the findings into editorial practice. Field notes on my blog. What I offer on my services page.
-
My work explores how AI affects creative judgment and the way people produce, evaluate and share knowledge, from cognitive offloading and bias to authorship, attribution and disclosure. The literature informs my approach.
-
The ground moves fast: models, search behavior, GEO/SEO, preferences on form (written and visual), regulatory landscape. Qualia Intel is my independent publishing project on AI, knowledge and society. It is where I follow the wider questions around AI, knowledge and society, working through them in public and checking ideas against the evidence. Some of that thinking eventually finds its way into Exavoice. Visual arguments on Instagram; essays on Substack (new). A selection of Qualia Intel work is featured here on Exavoice.
Research and narrative
-
Verification used to rest on independent sources agreeing. AI makes this much harder. Synthetic and derivative content can cite and repeat each other, and repetition can make unsupported information look like verified fact. Mistakes can happen in many ways: citations can be hallucinated, quotes misattributed, and AI may misrepresent people, things or claims by dropping a qualifier, widening the scope, turning possibilities into facts, or overgeneralising, etc. Because of this, the practice has to be stricter: primary documents, provenance before quotation, data checked at origin.
-
Interviews, proprietary data, original frameworks, direct observation and reporting bring in material drawn from people, evidence and the world beyond the model, the kind AI search increasingly values. Where more is needed, I contribute qualitative research and case studies, finding the strongest angle, the argument and the through line.
-
Article 50 of the EU AI Act introduces new transparency obligations for certain AI-generated or AI-manipulated content, including machine-readable marking and disclosure of deepfakes and AI-generated text on matters of public interest. Independently of these legal duties, recording sources, AI involvement, human review, and production history is increasingly becoming good editorial practice. I help organisations build both into clear editorial standards and workflows.
Visual articulation
-
AI systems increasingly retrieve and interpret information across formats, including text, images, charts and structured data, but not consistently. That makes coherence across modalities part of editorial practice.
-
Trained in documentary film and photography, I approach images as part of the argument, not decoration. The same principle carries across pieces, essays, reports, scripts and other formats: one concept, without losing the core message or visual grammar.
-
The method follows the work. Visuals may be commissioned for a project, developed with generative AI or sourced from existing material. The choice depends on purpose, rights, budget, visual identity and the organisation’s approach to AI.
Open tabs.
Everyone has thirty of them open at once. Here are the two I keep coming back to lately:
Exploring CHI 2026 conference papers: how interfaces shape trust, opinions, cognition and society at large.
GEO, AI-mediated discovery and the questions conventional search thinking leaves out. I keep a running notebook on those shifts in the GEO diaries.
And separately: what I am reading right now.
Latorre’s Un nuevo contrato social: Gobernar un mundo con inteligencias no humanas. Heard him speak recently: Link to the recording here.
Labatut, The MANIAC. On Von Neumann, intelligence and the limits of reason. Written in English as a second language, in prose that humbles.
LATEST BLOGS
Information ecosystem intelligence
•
Cross-domain creativity
•
Cultural awareness
•
Trust
•
AI-search readiness
•
Human-AI workflows
•
Information ecosystem intelligence • Cross-domain creativity • Cultural awareness • Trust • AI-search readiness • Human-AI workflows •
GET TO KNOW ME
80’s child. French-speaking Belgian. Madrid, Brussels, Luxembourg hopper.
Got my hands on a computer as a little girl thanks to my father.
Used it to play calculator games he coded for us. No epic tale about how I quickly learned to do the same as a breastfeeding baby.
Curiosity-driven. Books, screens, pen, camera – my pals.
PhD Political & Social Sciences
Digital content creation
Documentary Arts
Business Education
Paper trail.
Bernes, L.-A., Bousetta, H., & Zickgraf, C. (Eds.). (2017). Migration in the Western Mediterranean: Space, mobility and borders. Routledge.
Bernes, L.-A. (2014). Plastic border: Ceuta, behind and beyond the walls. International Journal of Migration and Border Studies, 1(1), 5–26.
Bernes, L.-A. (2013). Frontière(s), contrôle et jeu d'acteurs : Le cas de Ceuta (Espagne–Maroc / UE–Maghreb) [Doctoral dissertation, Université de Liège]. Political and Social Sciences, F.R.S.-FNRS research fellowship, CEDEM.
Bousetta, H., & Bernes, L.-A. (2007). Muslims in the EU: Cities report. Belgium. Open Society Institute.
EU research (junior position) — FP6 European Research Program: multicultural citizenship and anti-discrimination policy in Belgium (CEDEM, 2006–2008).
PhD — Political and Social Sciences, University of Liège (F.R.S.-FNRS fellowship, 2008–2013). Fieldwork on borders, borderlands dynamics, control practices and mobility. Advanced MA — International Relations & European Integration, Université de Liège. MA — Business Administration & International Management, HEC Liège.
Documentary and visual arts training — documentary filmmaking, LENS Escuela de Artes Visuales (2019–2020); film direction, ECAM Escuela de Cinematografía y del Audiovisual de la Comunidad de Madrid (2018); editing with Diana Toucedo and Fernando Franco; color grading (DaVinci Resolve); editorial photography with Sofía Moro (2020). Active street and documentary photography practice.
Editorial and thought leadership (client work) — 10+ years of editorial work and content creation (EN/FR, ES). Core territory: organizational Change Management (CM), leadership dynamics, digital transformation, and B2B communication.
Qualia Intel — AI, knowledge and society. Essays on Substack; visual arguments on Instagram.
EXAVOICE — Editorial strategy. Exavoice helps expert-led organizations build authority with audiences and earn visibility inside AI search through original, trustworthy content, GEO/AEO advisory, and information ecosystem intelligence. Explore services.
Multilingual content: French (native), English and Spanish.
Visual practice and hybrid workflows: Core visuals on this website are AI-generated and edited. Individual articles may also include photography and other archival material.
Let’s talk.
What are you working on?