A presentation by Pier Luigi Parcu on Artificial Intelligence, elections, media pluralism and media freedom at the European Artificial Intelligence Observatory April 2, 2019
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Artificial Intelligence, elections, media pluralism and media freedom
1. AI, elections, media pluralism and media freedom
Pier Luigi Parcu, CMPF
Brussels, 2 April 2019
EUROPEAN PARLIAMENT
EUROPEAN ARTIFICIAL INTELLIGENCE OBSERVATORY
2ND MEETING
2.
3.
4. 0 100 200 300 400 500 600 700 800 900 1000
ebay+Paypal
Ant Financial
Netflix
Tencent
Alibaba
Facebook
Alphabet
Amazon
Apple
Microsoft
Market capitalization of the biggest internet companies
worldwide, as of 31 March 2019, in billions of U.S. dollars.
Source:Ycharts
5. Technological challenges to quality of information
• The centrality of few Internet giants and the development of AI given by their
limited number and their global reach in the key functions of mediating interpersonal
communication and disseminating contents have serious implications for the quality
of information
• At the same time big data has become the driver of the digital economy
• The mix AI and big data is causing an acceleration in the disruption of many
industries and the traditional information industry is one of the most impacted. The
problem is that media have an externality of huge importance in the functioning of
democracies and particularly of electoral processes
6. Technological challenges to quality of information
• A specific threats for democratic processes derives from voters profiling and micro
targeting. In abstract, reaching with specific information the right target could be
useful and truly informative. But when information is driven by automated and
obscurely directed forces (bots) the existence of serious problem is clear. Should
social media platforms be allowed to self-regulate in their management of bots?
Should bots be required to disclose that they are not human and who is responsible if
they do not?
• Still limited scientific certainty about how widespread “information disorder” is
(Wardle and Derakhshan 2017), and on what is the dimension of the real impact on
individuals choices and behaviours
7. Technological challenges to quality of information
• Nonetheless, recent events have led many scholars to question the role of social
media seen as responsible for distributing disinformation (Allcott and Gentzkow
2017; Tambini 2017), using manipulative psychometric profiling (Cadwalladr 2017),
undermining authoritative journalism (Bell, 2018; Allcott and Gentzkow, 2017) and,
ultimately, jeopardizing the fairness and transparency of elections” (Tambini 2018)
• Even if scientific evidence is still incomplete (but it is accumulating), a principle of
caution would require preventive intervention. Moreover, besides disinformation,
also freedom of expression and plurality of information are heavily impacted by data
driven algorithms information and bots
8. Technological challenges to quality of information
Summarizing several new threats to information and fair electoral processes emerge:
1) The presence of only few gatekeepers and the disappearance of many traditional
and local media, may drive towards the excessive standardization and
homogeneity of the sources of news and qualified opinions, which negatively
affects the quality and variety of information. The situation is exacerbated by a
substantial absence of editorial control and editorial responsibility in the
distribution model of information typical of major internet players; aggravated
by lack of transparency and secrecy in the operating algorithms
9. 2) The polarisation of opinions created by the sociological and technological
dynamics, further impoverish the democratic dialogue and cause the exclusion
of middle ground and conciliatory occasions for debate. These developments
influence negatively how political consensus is reached and maintained, even in
otherwise democratic environments
3) In this “information disorder” increases the impact of hate speech, which has
become a common pattern of political propaganda, even in democracies,
targeting minorities, women, migrants, and so on
4) AI (in particular machine learning) cannot be safely trusted to counter
disinformation at this stage of knowledge and control without serious risks for
freedom of expression
10. Technological and political possible responses
1) Recognition that quality information is a public good and may require public
support
2) Imposition of rules about editorial control and editorial responsibility for
platforms
3) Technical empowerment of professional journalism
4) Support of media literacy, fact-checking and qualified actions of disinformation
contrast
5) Impositions of transparency rules for any kind of online political advertising
6) Imposition of disclosure rules for bots of the type imposed for advertising on the
media
7) Strengthening of competition rules regarding major internet players in the field
of information pluralism
11. Thanks you for your attention!
@cmpfeui
@FCP_eui
@FSRComsMedia