What is the algorithmic surveillance that the government wants to put in place?



Illustration of a person browsing the Internet. – fancycrave1

  • To detect threats, the anti-terrorism bill presented on Wednesday perpetuates the algorithm technique, which allows automated processing of connection data.
  • The device should be used to identify suspicious profiles unknown to the police services – as in Rambouillet – and not “necessarily linked to Islamist networks”.
  • “This mass surveillance will potentially analyze all the messages of the population (…) Who speaks to whom, at what time… We fear that the pretext of anti-terrorism will be deviated from its initial role”, estimates Quadrature du Net.

Will the fight against terrorism go through artificial intelligence? A few days after the Rambouillet attack, the government presented a new anti-terrorism bill on Wednesday, which perpetuates and extends the use of the criticized technique of algorithms to try to detect radicalized people, currently under intelligence radar. “It is a text which makes it possible to adapt to new threats that are less easy to detect and to rely on new tools linked to new technologies”, explained the Prime Minister, Jean Castex.

Of the thirty-five attacks foiled since 2017, “two” have been “thanks to the digital traces” left by their authors, underlined the Minister of the Interior, Gerald Darmanin. “The murderers of Samuel Paty and the parishioners of the Basilica of Nice communicated with their interlocutors only by encrypted messaging, Facebook and Messenger and not by telephone (…) by Internet, by encrypted messaging and social networks ”, insisted the Minister of the Interior.

What does the bill provide?

To detect threats, the anti-terrorism bill presented on Wednesday perpetuates the algorithm technique, which allows the automated processing of connection data. This practice, which was authorized for the first time on an experimental basis by the Intelligence Law of July 2015, has been operational with Internet service providers (ISPs) since 2017. The expiry of the experiment, provided for in the origin for 2018, has been postponed twice and is now due to end at the end of 2021.

The bill seeks to expand this system by extending it to web addresses (“URLs”). It also increases the authorized period for collecting computer data to two months, compared to just one today. Beyond that, these data are considered “dead” but may be kept for five years for the purposes of research and development of the artificial intelligence of the “black boxes” of the intelligence services.

The device should be used to identify suspicious profiles unknown to the police services – as in Rambouillet – and not “necessarily linked to Islamist networks”. Monitoring by algorithm allows for example to “detect unknown individuals from their exchanges on the Internet” or to identify “that an individual in France has had contact with an individual in western Syria”, detailed
on France Inter Laurent Nuñez, national coordinator of intelligence and the fight against terrorism.

How does algorithmic surveillance work?

“Black boxes” collect and analyze metadata (date of sending a message, location, device used, etc.). Equipped with algorithms spotting “weak signals” of radicalization or of a passage to the act, the “boxes” are installed at key places of the network with the hosts and the providers of Internet access (Google, Orange, Bouygues…). These are “automated processing”, “without collecting data other than those which meet the criteria” provided.

“The objective is to be able to collect, process, analyze and cross-check a large number of anonymous technical elements in order to detect low-intensity signals on the raw data which would testify to a threat to national security”, underlined the rapporteur of the bill, Jean-Jacques Urvoas. This involves an automatic and almost instantaneous analysis of the connections made on various Internet sites.

Why is this measure causing concern?

The bill, the final vote of which is expected by the government “before the end of July”, arouses an outcry from defenders of public freedoms. Many fear that the state’s use of algorithms will be extended to other areas, beyond the fight against terrorism. “This mass surveillance will potentially analyze all the messages from the population. (…) Who speaks to whom, at what time… We fear that the pretext of anti-terrorism will be deviated from its initial role ”, thus estimates
La Quadrature du Net, association for the defense of citizens’ rights and freedoms on the Internet

In terms of guaranteeing individual freedoms, the Minister of the Interior Gérald Darmanin explained “that listening to, monitoring a person detected by the algorithms” would be subject to “authorizations”, adding that parliamentary control was also planned. It is only a question of “applying to the Internet what we apply to the telephone” in terms of surveillance, he argued.



Source link