Note: this article was translated from French by an automated tool “Congratulations, Emmanuel Macron. Congratulations to French people for choosing Liberty, Equality and Fraternity over tyranny of fake news »Donald Tusk, President of the European Council, tweeted on the evening of May 7, 2017. It is true that the French electoral campaign was the unprecedented theater of attempts at destabilization and massive dissemination of inaccurate information, in the wake of a very agitated American campaign. Paradoxically or not, it is also Donald Trump who took charge of qualifying the phenomenon in the face of a most classic media, launching to the CNN correspondent, from his first press conference: " I'm not going to give you a question. You are fake news ».
Almost a year later, in a much less controversial fashion, Emmanuel Macron expressed concern, during his vows to the press, about "the irruption in the media field of false news, fake news "," Media which propagate them "," entire sites which invent rumors and false news which rank alongside the true ones ", and the threats that this phenomenon poses to democracy. And the President to announce a forthcoming bill to facilitate the fight against " fake news »During an election period.
It cannot be denied that such a prospect spontaneously arouses reservations, and not only because it calls into question a right as fundamental as freedom of expression. It illustrates, in fact, this reflex, after all very French, of responding with an additional law to a social problem, the magnitude of which in this case exceeds the French framework. The debate is of course only just beginning. We will content ourselves with briefly asking a few questions here.
1- Is it really necessary to fight against fake news ?
The limited impact of fake news on opinion
Ironically or not, the French President's announcement came just days after the publication of a major study by three political scientists on the US election (A. Guess, B. Nyhan, J. Reifler, Selective Exposure to Misinformation: evidence from the consumption of fake news during the 2016 presidential campaign). The authors combined analysis of web traffic in the months before and after the US election with the results of a poll of 2525 Americans. The fake news are defined there as "stories (stories) intentionally false, which can be verified to be false and which can mislead readers, ”a definition borrowed from economists Hunt Alcott and Matthew Gentzkow (see their study cited below). A website is considered fake news once it is established that he has published at least two false contents.
While some results of the study are unsurprising, others are more surprising (B. Carey, " "Fake News": Wide Reach but Little Impact, Study Suggests ", New York Times, January 2 2018). even though fake news were widely spread during the campaign - one in 4 Americans having seen at least one inaccurate content - they mainly concerned Donald Trump's camp. Not only 80 sites supported the current President on the 289 sites of fake news identified by the study, but nearly two-thirds (65%) of visits to these sites came from the 10% of voters most to the right of the political spectrum. Trump supporters were 3 times more likely to visit sites of fake news them supporters by Hillary Clinton.
The study shows, however, that the share of fake news in all information received by Americans has remained low: 1% for Clinton voters and 6% for Trump voters. They read, on average, only 5 fake news over a 5 week period. And if we judge by the sites then consulted by the interested parties, little false news was taken into account: only 10% of the readers then read other fake news. The study concludes that even the most frequent readers of fake news mainly interested in real news.
This conclusion is in addition to that of a study previously carried out by economists Hunt Alcott and Matthew Gentzkow (Social Media and Fake News in the 2016 Election, Journal of Economic Perspectives, vol. 31, numer 2, spring 2017, pp. 211-236) that only 8% of American adults believed in fake news, in a context where every American citizen has been exposed to at least one fake news during the campaign. Another study, published in March 2017, showed, from an analysis of social networks, that Americans generally obtain information from a plurality of media with the exception of the most conservative voters. However, she underlines that the most extreme media comments and themes influence other media, if only on the occasion of the fact checking (Y. Benkler, R. Faris, H. Roberts, E. Zuckerman, Breitbard led right-wing media ecosystem altered broader media agenda, Columbia Journalism Review, March 3, 2017, ).
The usefulness of the struggle
Is it necessary for all that to estimate, as it is titled the The New Yorker a few days ago, that the fight against fake news is useless (Masha Gessen, Fighting Fake News is Not the Solution, The New Yorker, January 4, 2018)?
We can doubt it. Despite these reassuring US studies, other analyzes highlight the worrying expansion of false or questionable information on social media. During the French presidential campaign, the English company Bakamo studied nearly 800 sites and 8 million links shared on French social networks between the 1er November 2016 and April 4, 2017 (The Role and Impact of Non-Traditional Publishers in the French Elections 2017). She found that 19,2% of the links shared referred to media that did not adhere to "journalistic standards" and expressed radical views, of which a fifth were clearly influenced by media owned by the Russian state. In addition, 5% of the links shared concerned conspiracy sites or sites referring to mythological or theological stories, again linked, for almost half of them, to Russia. Internet users sharing these sources have also been particularly active since sharing twice as many links as users consulting traditional sources. Another study by Oxford researchers found that a quarter of the political information shared on Twitter during the French election campaign was inaccurate (French social media awash with fake news stories from sources "exposer to Russian influence" ahead of presidential election, The Independent, April 22, 2017).
The fake news do not, moreover, exclusively concern electoral periods. For example, the recent decision of the Federal Communications Commission US to return to the principle of net neutrality was based on anti-neutrality comments posted on its site, of which it appeared that nearly 83% did not come from real users of the net (Millions of net neutrality comments were fakes, USA Today, December 6, 2017). In the field of consumption, a survey by the DGCCRF published in November revealed the very large number of false consumer opinions posted on the platforms (see DGCCRF, False consumer reviews on digital platforms, November 17, 2017).
2- The role of platforms
Is it the responsibility of the platforms to intervene first and foremost? Facebook, which studies have shown to be the main point of access to fake news, has set up a red sticker indicating "contested by third parties" for pages whose fact checker believe that they contain false information. To be identified as fake news, a publication must be qualified as such by at least two of these fact checker which are, in general, partner media (in France: AFP, BFMTV, L'Express, France Médias Monde, France Télévisions, Liberation, Le Monde and 20 Minutes). A pictogram is then displayed, indicating that the information is disputed. Internet users who are about to share questionable content receive a warning. In principle, users can find out more by clicking on a link to one of the media verifiers.
The effectiveness of this strategy is, however, disputed. A study by Yale researchers showed that the presence of the “third-party contested” pictogram only increased the number of Internet users doubting the veracity of the content by 3,7%. There would even be a counterproductive effect, causing users to believe unreported content more, even if it could not be verified (G. Pennycook, D. G Rand, " The implied Truth Effect: Attaching Warnings to a Subset of Fake News Stories Increases Perceived Accuracy of Stories without Warnings », December 8, 2017). In addition, the need to have the advice of at least two verifying bodies makes the operation complicated in countries where there are no or few independent and reliable organizations.
Facebook therefore recently decided to test a new strategy. Its algorithm now reveals, on a given subject, various articles, and in particular elements of fact checking, in the Internet user's news feed. It is to be hoped that this method will reverse the trend noted by the study by Guess, Nyhan and Reifler, which showed that it is mainly people who do not read fake news who read the articles of fact checking...
Mark Zuckerberg also announced on January 19, on his Facebook account, that the sources presented as a priority by the Facebook news feed would be the subject of surveys to establish the level of reliability that users attribute to them. Other measures exist: deactivation of fake accounts disseminating fake news, enhanced examination of sponsored content, information on the origin of advertisements (set up in Canada)…. While these efforts have so far borne some fruit (Facebook's global fight against fake news, CNN, May 9, 2017), are they sufficient? In the face of particularly viral and harmful fake news, eradicating the false information outright is a drastic but effective method. And relying, to do this, on a judicial decision guarantees the holding of an adversarial and fair debate. Going in this direction, however, amounts to entrusting the judge with the delicate task of determining what can and cannot be said.
3- Does the future law really threaten democracy?
The President announced that the provisions of the upcoming law will create a special regime, applicable during an election period, which will include increased transparency obligations imposed on platforms for each sponsored content. The identity of the advertisers and those who control them must be made public and the sums devoted to this content will be capped. In addition, a new summary action will be created, which will make it possible to have disputed content judicially removed, or even to close the user account from which it originated or to dereference the site or block its access. It is also planned to reform the regulation of television services and to rethink the powers of intervention of the CSA.
Among all these measures, it is, of course, the one which makes it possible to judicially obtain the deletion of a given content, or even the blocking of the site, which seems the most infringing on freedom of expression. Is ensuring a form of "policing of speech" by distinguishing between truth and falsehood really a role for the judge? Is it up to the law to repress an inaccurate statement on the sole ground of its falsehood?
Towards the imposition of an “official truth”?
Such questions are not new. They were raised, in particular, during parliamentary work relating to the so-called “Gayssot” law of July 13, 1990, article 9 of which introduced, into the law of July 29, 1881, an article 24 bis repressing revisionism (U. Korolitski , "Should we repress revisionism because it is false?", Political reasons 2016/3, n ° 63, p. 85-102). Some saw in this measure the imposition of an "official truth", even though "historical truth rejects any official authority", to use the words of Madeleine Rébérioux, who added "only one solution: to know and to make known" ("Against the Gayssot law", Le Monde, 21 may 1996, ). The same argument could be made today, even if the facts related by the fake news are not historical but contemporary: wanting to repress the fake news on the sole ground that they are false would lead to the imposition of an “official truth”, in defiance of freedom of opinion.
Seeing in the repression of a discourse on historical facts the imposition of an "official truth" constitutes, however, a form of confusion, as Yan Thomas has shown ("The truth, the time, the judge and the historian ”, Le Débat 1998/5, n ° 102, p. 17-36). If the nuance is fine, sanctioning the denial of certain historical facts (revisionism) does not exactly amount to asserting that the facts in question existed. Likewise, repress fake news does not consist in imposing an official discourse on reality but in correcting the situation caused by the harmfulness of this false information when it is proven. It is difficult to forget, for example, the deleterious effect on public opinion of the misinformation disseminated about the situation in Aleppo (see False images and propaganda of the Battle of Aleppo, the Decoders, March 20, 2017).
An already binding positive law
Should we really, in such circumstances, judge from the outset that the announced measures constitute a "threat to democracy" (A. Alemanno, " Macron's fake news law will threaten democracy », Politico, January 7 201)? Anticipate a text so liberticidal that it would be intolerable? If we had to go in this direction, we would certainly have to start by denouncing the positive law in force, since it already contains similar provisions.
French law already allows the deletion of certain content to be ordered judicially. Who has forgotten the withdrawal from the sale, decided in summary proceedings in January 1996 and then confirmed in substance, of the book “Le grand Secret”, written by François Mitterand's former doctor, Claude Gubler? It is true that this withdrawal (finally) earned France a condemnation by the European Court of Human Rights, which ruled in 2004 that the ban should have been limited in time (the book was finally reissued in 2005). The fact remains that the texts of French law currently applicable make it possible to request, urgently or in substance, the deletion or withdrawal of publicly accessible content, for all kinds of reasons: invasion of privacy, defamation, infringement, unfair commercial practice. , dissemination of personal data relating to a natural person and, more generally, any comments likely to cause harm or cause disturbance. The removal order can now be given directly to hosts, as expressly provided for in the law of 21 June 2004 on confidence in the digital economy (article 6 I.8).
There is, moreover, in French law, a powerful arsenal of criminal offenses, the reinforcement of which does not appear to be envisaged. Article 27 of the law of July 1881 on the press punishes with a fine of 45 euros " the publication, distribution or reproduction, by any means whatsoever, of false news, of fabricated, falsified or deceptive items (...) when, done in bad faith, it has disturbed the public peace, or has been liable to the to disrupt ". Inaccurate information relating to a named person may justify a conviction for defamation. Other criminal offenses target behavior occurring during an election period, such as, for example, article L.97 of the electoral code which punishes one year in prison and a fine of 15 euros. "Those who, with the help of false news, slanderous rumors or other fraudulent maneuvers, have surprised or deflected votes, determined one or more voters to abstain from voting ". Likewise, it " Any candidate is prohibited from bringing to the attention of the public a new element of electoral controversy at a time such that his opponents do not have the opportunity to answer it usefully before the end of the electoral campaign ”(Art. L48-2, electoral code).
In the light of positive law, it is ultimately the usefulness of the new law, rather than its liberticidal nature, which could be questioned. Adopt a text expressly providing for the possibility of ordering the deletion of fake news will simply make it possible to facilitate, by specifying the conditions, an action that is already possible in theory. It remains to determine what is meant by "false news" or fake news in order not only to avoid the courts being confronted with incessant demands but also to preserve the freedom of the press, which must be able to continue to do its information work. Will it be necessary to require the demonstration of an intention to deceive the reader or the awareness of the falsity of the disputed remarks, as retained, for their study, the aforementioned American researchers? Will the exercise of the action be restricted to election periods? Should we characterize the risk that the disclosure of erroneous information constitutes for society? Should false information relate exclusively to political facts or could it be any type of information, such as for example the false assertion of the toxicity of a product? for the purpose of harming its manufacturer? Many details will have to be given.
It appears, in any case, very premature, at this stage, to immediately see in the announced text a problematic attack on democracy.
4- How effective for French law when the platforms are transnational?
French law will obviously not be able to apply outside France. Sites based abroad may well ignore the injunction, even legal, to remove such or such content. This is why it is envisaged that the judge could order the blocking of access to a disputed site.
France is not, for the rest, not isolated in its fight against fake news, a concern shared by its European neighbors. The German NetzDG law (Netzwerkdurchsetzungsgesetz) obliges social networks with at least 2 million users to delete, within 24 hours of their notification, clearly illegal content published on their platforms, under penalty of incurring a fine of up to 50 million euros . For content whose illegal nature is less obvious, a period of 7 days, or even more, may be granted. This law, which does not provide for recourse for the benefit of the authors of deleted content, is widely criticized by defenders of freedom of expression. It mainly targets hate speech, defamatory content and incitement to violence already repressed by the German Penal Code: terrorist statements, speech defaming religions or religious or ideological organizations, Nazi symbols etc…. The fake news the most clearly illicit, in particular in what can be qualified as hate speech, are concerned. The case of Anas Modamani, a Syrian refugee living in Berlin who once posted a selfie taken with Angela Merkel is often mentioned. His photo was used to present him as the author of the terrorist attacks in Brussels and Berlin: it still circulates on social networks.
On the European Union side, the Commission has launched a public consultation until February 23, 2018. It has just announced that it will launch a program to fight against fake news before the summer, which will be based on the proposals of a group of around forty experts chaired by a Dutch academic, Madeleine de Cock Buning, and composed of representatives of platforms, researchers, journalists, and qualified professionals .
It is certain, on the other hand, that no similar law should emerge in the United States, where the freedom of expression guaranteed by the First Amendment to the Constitution is jealously protected.
