Three frequently confused terms
"Fake-news" has become a catch-all term mixing distinct phenomena. Three notions deserve a rigorous distinction, because they involve different actors, different responsibilities and different responses.
Disinformation. Deliberate diffusion of false or manipulated information, knowingly, with intent to deceive. It is intentional, often strategic, and conducted by an identifiable actor (political group, state, commercial interest). This is the strict, original definition of "fake-news".
Misinformation. Unintentional diffusion of false information by someone who believes it true. A relative forwarding a rumour on WhatsApp is misinformation, not disinformation. The distinction matters: the same false claim can be disinformation at the source and misinformation in the hands of someone sharing it in good faith.
Mal-information. Diffusion of truthful information removed from its context, distorted or used with harmful intent (privacy violation, reputation damage, harassment). The content is accurate, the intent is harmful.
Who produces disinformation
Organised disinformation is not a marginal activity. Several categories of actors devote significant human and financial resources to it.
State actors. Some states fund influence operations targeting foreign public opinions: electoral interference, diplomatic destabilisation, justification of a military policy. These operations rely on state media, dedicated Telegram channels and networks of inauthentic accounts on social platforms.
Lobbies and commercial interests. Petrochemical, agri-food, pharmaceutical, technology industries: any sector whose profitability depends on public perception has a direct interest in shaping the debate. Historically this happens via biased studies, front associations and paid opinion leaders.
Troll farms. Commercial firms employing operators to mass-produce content, comments and fake accounts on behalf of clients. The service is billed by reach (views, shares) or by campaign duration.
Individual scammers. Independents producing sensational or fraudulent content in a purely opportunistic logic of ad monetisation or direct scam (fake crypto investments, fake miracle products, fake petitions for data harvesting).
Most exposed domains
Four domains concentrate a large share of disinformation campaigns, because they combine emotional, financial and political stakes.
Electoral politics. Pre-election periods: manufactured scandals, manipulated quotes, fake polls, fabricated statements attributed to candidates. The calendar is tight, virality outpaces verification, and corrections always arrive after the vote.
Health miracles. Miraculous supplements, natural cures for serious diseases, vaccine denial, theories about treatments "hidden by the industry". This register exploits real distress in the face of illness and distrust toward medical institutions.
Finance and cryptocurrencies. Fake guaranteed-return investments, fake trading experts, disguised Ponzi schemes, crypto theft via fake platforms. The sector massively attracts fraud because transactions are fast, hardly reversible, and technical complexity serves as a screen.
Climate scepticism and environment. Systematic doubt cast on scientific consensus, exaggeration of climate-policy costs, fabricated controversies around devices (wind turbines, electric vehicles). A domain historically funded by documented industrial interests.
Recurring techniques
Five techniques appear in almost all disinformation campaigns, regardless of domain or actor.
Cherry-picking. Selecting only data that confirms the defended thesis, systematically ignoring contradicting data. Statistically inadmissible, but visually convincing for an uninformed reader.
Fake experts. Promoting personalities labelled as "experts" with fabricated, unrelated or out-of-specialty credentials. A dermatologist commenting on vaccination, an economist commenting on climatology: legitimacy is implicit, real expertise is absent.
Astroturfing. Simulating a spontaneous citizen movement ("grassroots") while in reality orchestrating a paid campaign. Fake citizen accounts, fake testimonies, fake petitions: the goal is to feign a consensus that does not exist.
Audio and video deepfakes. AI-generated videos or voice recordings in which a public figure appears to say things they never said. The technology is now accessible, the cost negligible, detection hard without specialised tools.
Bot networks. Thousands of automated accounts artificially amplifying content (coordinated likes, shares, comments) to feign organic virality and trigger algorithmic promotion.
Verification methods
Verification is not the exclusive province of journalists. Three practices structure a solid critical analysis accessible to any reader.
Trace back to the primary source. An article citing "a study" without a direct link, or "a report" without verifiable reference, should be treated as unsourced until proven otherwise. The primary source (scientific publication, official document, original statement) should be reachable in two clicks.
Identify your own confirmation bias. Information that perfectly matches our convictions deserves more rigorous verification, not less. The natural reflex is the opposite: we easily accept what suits us and question what challenges us. This bias is precisely what disinformation campaigns exploit.
Consult independent fact-checkers. Specialised newsrooms (AFP Factual, Snopes, EU vs Disinfo, PolitiFact, antiphishing.ch for fraud) publish detailed analyses of recent rumours. A keyword search on their site is usually enough to clarify a viral claim.
Further resources
For verification resources and fact-checking sites, see our Useful links page (category "Anti-fraud and reporting"). Our Disinformation / Clickbait page specifically addresses the attention-manipulation mechanics on social media and the cognitive biases to which young audiences are particularly exposed.