Algorithms play a fundamental role in current information. In the past,

traditional media managed the production, verification and communication of

news. In online information, instead, it is algorithms used by the platforms that are

relevant. Algorithms are able to sort the information that appears on the users'

screen. This order of news, however, is not the result of an objective and neutral

search: instead, algorithms select the information to be proposed to individual

information role that traditional mass media had in the past: in fact, they would no

longer be able to select and control information or to deny false news (Levi 2018,

Baker 1998). This is also because citizens tend to get more and more informed

through the Net and less and less through traditional media. Indeed, many citizens,

especially younger ones, inquire only through the Net.

Secondly, the users, through profiling operated by algorithms, are at risk of

receiving only that news that reflects their orientations, opinions and prejudices.

On the other hand, users also tend to search online for information that reflects

their way of thinking and share it with other users. Furthermore, algorithms, in

addition to selecting the news to be proposed to each user, would also produce

but are also able to influence users, who tend to believe it, because it is truthful,

but above all because it is in line with their ideas. Moreover, the average user tends

hoc fake news, for particular purposes. The purposes can be: limit the right to be

correctly informed; orienting public opinion; discredit politicians and governments

and their actions and, consequently, condition the voters’ vote. This fake news,

artfully created for specific purposes, is potentially capable of producing serious

effect’ contribute to amplify false news especially among like-minded people. On

the other hand, the phenomenon of ‘confirmation bias’, leads users to believe not

so much in the true news but in that which confirms their opinions and their

prejudices. Consequently, these users will tend to share information especially

of freedom of information like the basis for the formation of a public opinion that

leads to a conscious exercise of civil and political rights.

democracy. Algorithms are increasingly present in policy decision-making

processes but this should not replace democratic systems based on indirect

democracy.

they receive online is selected on the basis of their profiles but believe that it is

neutral. In any case, users who receive selected information, based on a profiling

that could influence their behavior, ignore the methods and operating criteria used

there is a danger for democracy that would justify regulatory intervention by

States, some proposals for measures that can be taken to deal with this phenomenon

will be put forward.

between profiling online users, through the algorithms, the fake news created

specifically to manipulate public opinion and the dangers for democracy. Next, a

literature review will show how the recent doctrine has focused on these issues and

many authors have highlighted the risks to democracy due to the spread of fake

democratic electoral processes. On the other hand, the role of algorithms in the

policy decision process will be analyzed and the risks for representative

democracy will be assessed. The framework that will be described in the work,

will lead to a belief that a regulatory intervention is necessary to limit the creation

and dissemination of that news, specifically created for the purpose of influencing

the democratic systems. On the other hand, the hypothesis of regulatory

interventions, that limit the use of platforms that involve users in the policy

decision process, will also be evaluated, because the mechanisms of operation of

algorithms are obscure and many citizens would be excluded. Then, the possible

and the possibility of denying them, through debunking tools, rectifying them or,

sometimes, removing them, after an evaluation by a judicial authority; enhancing

the Media Literacy of network users to develop in them a critical spirit towards the

news. Then, the effectiveness of the various proposed measures will be assessed,

also on the basis of experiments conducted in some countries and researches

conducted, in this regard, especially in the United States. Finally, we will conclude

our results arguing that a single measure, provided for by state regulations, would

not be able to limit the phenomenon of fake news and to make the algorithms’

operating methods more transparent. In conclusion, therefore, taking into account

each measure considered has limits, it will be argued that a regulation that

Research on the impact of algorithms on democracy has started in the

international arena in a relatively recent period. Some of the work in the last years

is summarized below.

(2006), Hodas and Lerman (2012), McNair (2017), Nematzadeh et al. (2017) and

Riva (2018).

users tend to believe more in false news on the Net than in real news spread by

traditional mass media, because they no longer trust them.

formation of public opinion and democracy, research has been conducted by

Morozov (2011), Panarari (2017), Papa (2018), Cuniberti (2017), Caravita (2018),

Manetti (2018), Pitruzzella (2017), Costa (2016), Borello (2017), Pinelli (2017),

West (2017), Syed (2017), Balkin (2018), Boccia Artieri and Marinelli (2018),

Gurumurthy and Bharthur (2018), Mezza (2018), Smith-Roberts (2018).

platforms to a black box: in the same way as black boxes the platforms collect our

reinforcing them. Other studies on the topic have been conducted by Zuiderveen

Borgesius et al. (2016) and Pitruzzella (2017).

(2017), Bassini and Vigevani (2017) and Gallone (2019) have studied the possible

effects of fake news on democracy.

can influence and manipulate the consent of the voters and, consequently, their

vote. Frosini (2017) has a different opinion. Huighe (2016), Berghel (2017),

(2011), Bakshy et al. (2015), Rožukalne (2015), Huighe (2016), Silverman (2016),

Quattrociocchi and Vicini (2016), Quattrociocchi and Vicini (2018), Giglietto et

al. (2019), Iacoboni (2016), Gili and Maddalena (2017), Spohr (2017), Grassegger

and Krogerus (2017), Bistagnino and Fumagalli (2018), Del Vicario et al. (2019),

Perucchietti (2018), Visco Comandini (2018) and Reinemann et al. (2017) have

highlighted the changes in online political information.

Furthermore, recent studies have dealt, more specifically, with the tools to

removal systems of fake news.

De Keersmaecjer and Roets (2017), Royster (2017) and Jolls and Johnsen

(2018) believe that Media Literacy is ineffective.

Nyhan and Reifler (2010), Nyhan (2017) and Pennycook and Rand (2017)

believe that fact checking and warning systems are ineffective.

The studies of Zollo et al. (2017) and Baron and Crootof (2017) showed that

debunking is ineffective.

the considerations made would not be understandable. For this reason, the results

of researches carried out in recent years have been examined, which concern the

analysis of the facts and practical implications related to the impact of algorithms

on democratic processes. On the basis of this researches, also based on the

empirical experiences that have taken place, in the United States and in European

countries, it will be possible to argue that platforms that use algorithms are

potentially able to produce negative consequences on pluralist and representative

democracy.

The methodological approach is the legal one and the solution proposed to

deal with this phenomenon is based on regulatory intervention.

European Union, have revealed the limits of these measures.

Therefore, the expected result of the research is not to identify the best

solution to deal with the problems covered by the study but, rather, to demonstrate

the need for a regulatory intervention that provides for a mix of measures.

Measures that, even if they cannot completely solve the problem of the impact of

algorithms on democracy, can, at least in part, limit its negative effects.

those concerning politics, in general, and election campaigns, in particular. Thanks

Leveraging on profiling, politicians tend to give more prominence to news that can

give them greater visibility and to arouse empathy and emotions in users

(Ratkiewicz et al. 2011, Silverman 2016, Shao et al. 2018). The objective of the

political information, in this case, is no longer to inform about the activity of

politicians or parties, but to obtain consensus and discredit opponents and their

actions, through the creation of fake news and ad hoc reality (Keyes 2004, Gili and

Maddalena 2017, Grassegger and Krogerus 2017, Pitruzzella et al. 2017,