Legal framework

Editorial Ethics & AI

Our detailed code of ethics. Updated: 23 May 2026.

1. Independence

BXLpositive.news is published by the non-profit organisation BXLOVE NEWS PRESSE IA ASBL, with no commercial shareholders, without public subsidies tied to a specific editorial line, and with no exclusive contracts with any political party or institution. All financial resources (donations, patronage, partnerships, advertising) are tracked and published in our annual report.

2. Honesty in reporting

  • No information is published without a verifiable, cited, and linked public source.
  • No reconstructed quotes: only verifiable statements (from video, press articles, or press releases) are placed in quotation marks.
  • No AI “hallucinations” are tolerated: any fact that cannot be sourced is removed before publication.
  • A clear distinction is made between facts, analysis, opinion, and sponsored content.

3. Respect for individuals

  • Presumption of innocence: no individual is named in ongoing legal proceedings without a court ruling.
  • Protection of minors: no child’s face or identity is published without written parental consent.
  • Right to be forgotten: personal data can be removed in accordance with Article 17 of the GDPR.
  • Image rights: removal is examined upon a simple, reasoned request, as a priority when urgency is flagged.
  • No stigmatisation based on origin, religion, orientation, gender, disability, or social precarity.

4. AI Transparency (AI Act, Art. 50)

  • AI assistance is indicated in the byline of the articles concerned.
  • The model used (e.g. google/gemini-2.5-flash) is identifiable in the metadata.
  • AI-generated images are marked “AI image” (integration of a provenance standard such as C2PA is under consideration).
  • No deepfakes of living real people.
  • No synthetic voices imitating an identifiable person.
  • A recorded human review is required for every new publication; earlier archives are not validated retroactively (actual status in the /fact-check log).

5. Editorial workflow (5 steps)

  1. Detection — Our AI agents rely on identified public sources from Brussels (BX1, Brussels Times, Bruzz, RTBF, Le Soir, official press releases, local press).
  2. Positive angle filter — Any information where a positive angle would feel artificial is discarded.
  3. AI rewrite — In the agent’s style, with the source cited and a direct link provided.
  4. Human verification — Checking sources, facts, tone, and rights.
  5. Publication — Visible AI notice, right of reply available.

6. Conflicts of interest

Any institutional, event, or commercial partnership is indicated at the top of the article (with the notice “Sponsored content” or “In partnership with”). Members of the non-profit organisation (ASBL) may not cover an organisation in which they have a direct interest.

7. Corrections & right of reply

Any substantial correction is noted at the bottom of the article, indicating its date and nature. See right of reply, rectification & removal.

8. Comment & social media moderation

We moderate content in accordance with the Belgian penal code (incitement to hatred, defamation, harassment) and our charter (no bashing, no ad hominem attacks). All moderation is carried out by humans, is documented, and can be appealed.

9. European commitment

We take as a reference, without certification or external audit, the rules of the European Media Freedom Act (EU Regulation 2024/1083), the Digital Services Act (EU Regulation 2022/2065), and the EU Artificial Intelligence Act (EU 2024/1689). See our AI Act page.

10. Appeals

In the event of a persistent disagreement, you may contact the Conseil de déontologie journalistique (CDJ), the Autorité de protection des données (APD), or the relevant Belgian courts. See our FAQ.