Phasing Out Journalism in Favour of Crowdsourcing
Meta's Oversight Board, an independent body consulted on the tech giant's content moderation policies, has formally warned corporate leadership against a concerning shift. At issue: trials conducted across sixteen Latin American countries to introduce a "community notes" system, modelled on the mechanism used on X. This system tasks regular users with drafting and evaluating contextual corrections on viral posts, raising fears of a gradual retreat from paid partnerships with professional newsrooms.
This warning resonated immediately with the LatamChequea network, which brings together nearly fifty investigative journalism and fact-checking organizations across South America. According to these information professionals, delegating the arbitration of factual reality to participatory voting in highly polarized political environments directly exposes platforms to coordinated manipulation campaigns and heightened partisan pressure.
This impulse to replace the rigorous analysis of qualified journalists with unpaid, algorithmic moderation raises a fundamental public governance question: can factual truth truly emerge from a statistical consensus among users?
Structural Flaws of Participatory Fact-Checking
Understanding platforms' interest in participatory models requires examining their economic and architectural logic. Maintaining global partnerships with news agencies and investigative bureaus represents a substantial financial commitment, while repeatedly putting the platform in an uncomfortable position as a political referee. Shifting to community notes aims to offload this moral and operational burden onto users themselves.
Technically, these systems typically rely on bridging algorithms. For an explanatory note to become visible beneath a post, it must earn the approval of users who have historically held opposing viewpoints. While the approach seems appealing in theory, scientific research highlights its inherent limits. Researchers at the Massachusetts Institute of Technology demonstrated that while crowds can identify blatant falsehoods, they struggle immensely with subtle manipulation, complex scientific data, or unfamiliar regional contexts.
Even more paradoxically, as noted by the German newspaper Die Zeit, the vast majority of community notes published on social media rely directly on reporting and articles produced by professional journalists. Defunding investigative teams on the pretext that the crowd can replace them ultimately dries up the documentary source material that same crowd needs to support its claims. Without verifiable primary sources, statistical consensus no longer measures factuality; it merely reflects the mobilization strength of organized groups.
Documentary Rigour Versus the Dictatorship of Engagement
In response to the pitfalls of platforms driven by attention and emotional reactions, pursuing informational reliability demands a return to the foundational principles of documentary methodology: source traceability, a diversity of viewpoints, and insulation from engagement metrics.
This structural distinction is embedded directly within the ProductivIA ecosystem architecture. The Actualité application, which delivers regional and national news briefings, operates on an explicit aggregation model of verified journalistic feeds. Rather than leaving event visibility to community amplification or emotion-driven algorithms, the platform transparently shows the institutional and media origin of each dispatch. Users know exactly which newsroom produced the report, when it was published, and how the facts were gathered.
This same methodological standard guides the Base documentaire application, which supports organizational knowledge and AI-assisted decision-making. To neutralize hallucinations in large language models, the platform relies on retrieval-augmented generation, commonly known as RAG. This technology does not fabricate content or rely on popularity votes: when a query is submitted, the system performs a semantic vector search across archival records, audited reports, or regulatory texts held in a secure repository.
The conversational assistant extracts relevant excerpts and formally cites the archived source. This direct traceability allows users to verify every element of an answer by consulting the original file stored transparently in the Nuage application, under local sovereignty and in strict compliance with Law 25. Information reliability never depends on the noise of an online crowd, but rather on an uncompromised chain of textual evidence.
Rethinking the Role of Expertise in the Digital Age
Meta's pilot highlights the fragility of an approach to knowledge that outsources truth entirely to the public. As generative algorithms and social platforms accelerate the spread of ambiguous or intentionally misleading content, the value of information hinges more than ever on the transparency of how it was verified.
Do public institutions, educational networks, and businesses really benefit from abandoning methodical verification in favour of transient reputation metrics? Preserving a healthy information environment will likely require reaffirming the indispensable role of professional journalism, backed by computing architectures built to index, preserve, and faithfully cite written records.