The opinions published online by consumers can be an excellent barometer to estimate the quality of a product or service for sale. While it is not uncommon to refer to them to guide an election, these assessments are not always reliable. It is within this framework, and to fight against comments, ratings or other fraudulent valuations in the public domain, which the General Directorate for Competition, Consumption and Fraud Prevention (DGCCRF) is about to make available to its researchers, “Polygraph”, a tool to detect them
A Bercy decree, examined on December 15 by the National Commission for Informatics and Liberties (CNIL), should allow the implementation of this system, the development cost of which was estimated in 2019 at 300,000 euros over two years, which should make it possible for specify the professionals guilty of disseminating suspicious notices, which is considered a crime, reported this Tuesday, January 3, L’Informé.
Raw data exploitation
In detail, according to the Ministry of Transformation and Public Function, Poygraphe will retrieve and use a lot of “relevant data” on the platform in question (site, phone number, location, etc.) as well as on the taxpayer (name or nickname, identifier in the platform, URL of your page, etc.). In addition, elements related to the reviews (text of the comment, text of the possible response to the review, etc.), or connection data and other event logs (add or delete an account, nature of the requests made, content modification by administrators, etc.).
The algorithm then goes“identify suspicious comments using various indicators of suspicion defined with experienced investigators in the matter”, indicates the Ministry of Transformation and Public Function. The last phase consists of “the visualization of the results in the form of an interface for researchers”. To manage the use of these data, the DGCCRF wishes to be able to keep them for up to 6 months from their collection, in addition, the right of Internet users to oppose these treatments will not apply.
The Constitutional Council could veto
However, this device could be challenged by the Constitutional Council. The latter had estimated, within the framework of an article of the 2020 finance law that authorized tax services to collect and use data made public on online platforms, such as Facebook or Instagram, in order to detect the tax domiciliation fraud, that “data that reveal the alleged racial or ethnic origin, political opinions, religious beliefs or philosophical or union affiliation of a person, genetic and biometric data and those related to health and life or sexual orientation.
However, the use of online reviews could potentially allow the DGCCRF to develop a personal profile based on all the comments of an Internet user. For example, the organization might be able to infer a person’s religious opinion if, for example, they left a review on a halal butcher’s site or even learn about their health status after seeing a notice left on the website. from an optician. .
Source: BFM TV
