DEMOCRACY & DISINFORMATION: A TURN IN THE CURRENT DEBATE Anja - - PowerPoint PPT Presentation

11 OCTOBER 2019

DEMOCRACY & DISINFORMATION: A TURN IN THE CURRENT DEBATE

Anja Bechmann, Professor – Media Studies, Aarhus University, Denmark Ben O’Loughlin, Professor – International Relations, Royal Holloway University of London, UK #battleforthetruth & @_KV AB

11 OCTOBER 2019

n ), t

PRESENTATION

u

Why a turn in the debate? Conceptual discussion

u

Evaluation of existing recommendations

u

Recommendations for academics and platforms

u

Flandern particularities

u

Recommendations for political leaders and journalists

11 OCTOBER 2019

n ), t

1.

Conceptual turn

2.

Combat strategic turn – already taking place

3.

Data Access strategic turn

A TURN IN THE DISINFORMATION DEBATE?

11 OCTOBER 2019

n ), t

“false, inaccurate, or misleading information designed, presented and promoted to intentionally cause public harm or for profit” (EU Commission report, 2018) To a broader notion covering information disorders focusing on impact including: Misinformation (e.g. vaccine, immigration and climate change debate), hostile sentiments, ”fake news”, and disinformaiton disregarding intentionality

DISINFORMATION

11 OCTOBER 2019

n ), t

1)

it seeks to establish what we know and what we do not know about disinformation.

2)

it identifies what from a research point of view specifically it is important to pursue in future studies.

3)

it identifies policies and actions that should follow from the existing knowledge that could address the current dilemmas faced by democracies.

OBJECTIVES OF THE REPORT

11 OCTOBER 2019

n ), t

• Most conflicting discussion point due to conflicting values such as privacy vs. Safeguarding democracy, freedom
• f expression/mechanical neutrality vs. Censorship
• Good faith research soft law solution has been negotiated through HLEG work within EU
• This has resulted in e.g. social science one (good faith access), ad transparency archive and new call including

CrowdTangle access

• Lack of scalable solutions for the whole community
• Lack of freedom of science with Social Science One as intermediary instead of complying to EU regulation and

ethical conduct – do platforms pose questions through available data structures?

• Lack of focus on end-to-end encryption versus ”community” communication space knowledge as central for

democracy and societal values (compliant to regulation)

#1 ENHANCE TRANSPARENCY OF THE DIGITAL INFORMATION ECOSYSTEM

11 OCTOBER 2019

n ), t

Everyone can agree on this but stakeholders do not agree on what it is Data access and evidence-based research are prerequisites to make solid media literacy ”data literacy is power for social change”?

#2 PROMOTE MEDIA AND INFORMATION LITERACY

11 OCTOBER 2019

n ), t

Huge disagreement of specific initiatives in this field. Research-driven perspectives: E.g.

1.

reactance theory suggests that interfacing flagging would create more circulation,

2.

transparency in code of practice create negotiation ground in participatory democracy ideals

3.

concept and degrees of disinformation to whom in what context? -> partisan content, should we make a space for optimizing user time an elightment project? Or focus on EU/national supported media with must- carry obligations?

4.

International/regional Standardized data models open sourced for research/fact-checking

5.

Fact-checking a moving target due to socio-technological (tech+skills) developments and network structure e.g. outlier/disinformation detection and automation and/or crowdsourced biases – good enough signals to justify censorship?

#3 DEVELOP TOOL TO ENHANCE THE EMPOWERMENT OF USERS AND JOURNALISTS AND FOSTER POSITIVE ENGAGEMENT

11 OCTOBER 2019

n ), t

How sustainable is the economy of content providers in a long-term perspective? Existing studies suggest a steap decline going from analogue-digital-mobile Is this forcing media to produce more clickbait/bad journalism in third party spaces in proprietary spaces? Should we ”safe/factcheck” those spaces if business model is not aligned with democratic values and basic democratic goals of enlightment? How do we make a sustainable EU solution to create alternatives to US and China based solutions? That are not acquired by non-EU companies – big fives?

#4 SAFEGUARD THE DIVERSITY AND SUSTAINABILITY OF THE EUROPEAN NEWS MEDIA ECOSYSTEM

11 OCTOBER 2019

n ), t

• In order not to be vulnerable to political influence
• Access to (good) data is a prerequisite for evidence-based research
• Lacking studies at scale and across platforms and across physical and digital spaces

(contextual analysis)

• Lacking coordinated and standardized deeplink lists instead of source coding with

conflicting data models in order to compute at scale

#5 CONTINUOUS RESEARCH ON THE IMPACT OF DISINFORMATION IN EUROPE

11 OCTOBER 2019

n ), t

Short-term:

u

Implement a scalable data access solution that favor data access control rather than limit amount of data units and data types

u

Develop high ethical standards applying to the European context in collaboration with European research communities

u

Making data available in a format that allows for academic researchers to pose own research questions in the belief of academic freedom

RECOMMENDATIONS FOR PLATFORMS

11 OCTOBER 2019

n ), t

Medium-term:

u

Provide specific and differentiated APIs for (i) journalists and NGOs, and (ii) university-based academics in order to meet different GDPR concerns and societal interests Long-term:

u

Provide researchers with stable access to social media data in safe, controlled spaces that reduce privacy risks to users whose data is analysed, preferably outside platforms and across platforms.

u

Making internal research with a focus on to what extent social media data can be made non- identifiable from the beginning without compromising the need for opening up to academic researchers to pose research questions in the interest of society and with legal and ethical clearance

RECOMMENDATIONS FOR PLATFORMS

11 OCTOBER 2019

n ), t

Short-term:

u

The academic community needs to push for data exchange solutions that acknowledge different kinds of research questions in the support for academic and scientific freedom for the benefit of society

u

Academic communities and national foundations should support and encourage this topic to be conducted in collaboration across fields and across countries in order to recognize this as an international challenge Medium-term:

u

More research is needed in order to move away from a content-centric research agenda in the context of disinformation and instead move towards circulation at scale and across platforms

u

More experimental research is needed in order to measure effects of fact-checking, media literacy, reconfigurations of algorithms as well as networks in order to slow down or prevent circulation of disinformation and hostile sentiments

u

More work is needed in order to understand effects between disinformation exposure and voting behavior

RECOMMENDATIONS FOR ACADEMICS

11 OCTOBER 2019

n ), t

u

More research is needed in order to connect research on disinformation with larger questions of polarization and trust as such causal connections are sparse in existing studies – we need more studies to back this up and especially outside UK and US in small countries with high education level such as the Flemish region. Long-term:

u

Negotiating solutions for data access, processing and storage that cuts across individual academic incentives for the benefit of the general knowledge level and skillsets in our research communities for researchers at all levels both within small countries and larger regions

RECOMMENDATIONS FOR ACADEMICS

16 SEPTEMBER 2019

11 OCTOBER 2019

n ), t

Mainstream media = high trust, renewed revenue and agenda-setting online But anxieties:

Ø success of Vlaams Belang – and journalists give a platform to winners Ø fear of polarisation between paying / non-paying news consumers Ø worries about ‘normalisation’ – should Sceptre.be and Doorbraak.be be reported?

THE FLANDERS CONTEXT

11 OCTOBER 2019

n ), t

Vaccari (2017): High trust means high susceptibility to disinformation EU Commission research shows:

Ø Belgian citizens least likely across whole EU to say false news is a problem in their country Ø Belgian citizens express very low confidence in their ability to spot false news Ø They say: leave it to the government or news editors to stop false news

… social fragility?

A FRAGILE ELECTORATE

11 OCTOBER 2019

n ), t

Must now perform fact-checking in real-time to satisfy journalists Feel the Flanders public is more patient and will trust they will verify facts eventually Policymakers expressed no sense of being in an ‘information war’ – digital is an opportunity to promote Flanders and Belgium. Journalists and think tank staff – No: Brussels is a symbolic target! Diplomats a resource overseas to spot conspiracies and false news about Belgium

FLANDERS POLICYMAKERS’ VIEWS

11 OCTOBER 2019

n ), t

Short term:

u

Independent (high level) statement from the EU Commission announcing that data exchange does not violate privacy and GDPR

u

Political parties must agree to not use bots, sock puppets or other third-party techniques that hide identity in order to spread content

u

Political parties must continue to agree to disclose budget, commercial partnerships and content of political campaigns

u

Support building debunked lists on a story level in an open source solution with international data model format that encourage crowdsourcing

u

Political leaders must recognise the signal value of taking false content down – to deter those who would break the democratic norm of informed citizenship

RECOMMENDATIONS FOR POLICYMAKERS

11 OCTOBER 2019

n ), t

Medium term:

u

Political leaders must seek consensus on how to understand the tradeoff between privacy and transparency. They must be willing to enforce this and support this balance to external actors (the US, China, and so on) because this is a global phenomenon transgressing national borders.

u

Law and actions taken in Europe must be based on EU regulations, not US regulations.

RECOMMENDATIONS FOR POLICYMAKERS

11 OCTOBER 2019

n ), t

Long term:

u

EU Commission to provide non-financial support to establishment of a data exchange solution where data is stored outside and across platforms. Further work needs to be done in order to find optimal safe space solutions like DNA registers and other health data, and to create practices for private company data to be stored this way.

u

Politicians should create stronger support for research in sustainable tech and media landscapes that can provide alternatives to platforms operating under American values and commercial incentives. Such alternatives would ideally operate in continuation of strong press and public service traditions that favour democratic values, e.g. independent regulators.

RECOMMENDATIONS FOR POLICYMAKERS

11 OCTOBER 2019

n ), t

Short term:

u

Continue and support the norms of severe 'call outs' for journalists who reprint disinformation, either by other journalists/editors (as a new norm) or through industry bodies. Medium term:

u

Educate journalists in how platforms work, how data science works, and thus how to produce scientifically-informed reporting about platforms and using platform data, e.g. via MA/MSc journalism programmes or secondments/exchanges.

RECOMMENDATIONS FOR JOURNALISTS

11 OCTOBER 2019

n ), t

Long term:

u

Build a culture of responsibility in which drawing attention to disinformation is done cautiously and reproducing disinformation is a severe norm violation This would counter the 'escalating' and 'normalising' dynamic that journalist reporting of radical/extreme views has, making them mainstream and just encouraging even more extremism. Example

Ø In Northern Ireland during the conflict journalists on all sides agreed not to report violence

sometimes, because such reporting just encouraged revenge attacks and more violence. The journalist norm was “sensationalism costs lives' -> they collectively agreed to de- sensationalise their news.

RECOMMENDATIONS FOR JOURNALISTS

11 OCTOBER 2019

n ), t

Basic problem: We cannot research a phenomenon of historical importance, leaving an entire debate to be driven by assumptions and perceptions. Need to be clear about who is doing what at what level (regional, national, EU, global) Continued anxiety about the need for a narrative supporting liberal democracy - the earlier three (conceptual, strategic, data access) plus a narrative turn.

TO CONCLUDE

16 SEPTEMBER 2019

Ben O’Loughlin - contact: ben.oloughlin@hol.ac.uk Anja Bechmann - contact: anjabechmann@cc.au.dk

THANK YOU