We use cookies to give you the best online experience. By using our website you agree to our use of cookies in accordance with our cookie policy.

Council of Europe Report: Recommendations

coe_recommendations

About the report | Download the report

What could technology companies do?

  1. Create an international advisory council. We recommend the creation of an independent, international council, made up of members from a variety of disciplines that can (1) guide technology companies as they deal with information disorder and (2) act as an honest broker between technology companies.
  2. Provide researchers with data related to initiatives aimed at improving the quality of information. While technology companies are understandably nervous about sharing their data (whether that’s metrics related to how many people see a fact-check tag, or the number of people who see a ‘disputed content’ flag and then do not go on to share the content), independent researchers must have better access to this data in order to properly address information disorder and evaluate their attempts to enhance the integrity of public communication spaces. As such, platforms should provide whatever data they can—and certainly more than they are currently providing.
  3. Provide transparent criteria for any algorithmic changes that down-rank content. Algorithmic tweaks or the introduction of machine learning techniques can lead to unintended consequences, whereby certain types of content is de-ranked or removed. There needs to be transparency around these changes so the impact can be independently measured and assessed. Without this transparency, there will be claims of bias and censorship from different content producers.
  4. Work collaboratively. Platforms have worked together to fight terrorism and child abuse. Slowly, collaboration is also beginning to happen around information disorder, and we encourage such collaboration, particularly when it involves sharing information about attempts to amplify content.
  5. Highlight contextual details and build visual indicators. We recommend that social networks and search engines automatically surface contextual information and metadata that would help users ascertain the truth of a piece of content (for example, automatically showing when a website was registered, or running a reverse image search to see whether an image is old). The blue verification tick is an example of a helpful visual indicator that exists across platforms. We argue that technology companies should collaborate to build a consistent set of visual indicators for these contextual details. This visual language should be developed in collaboration with cognitive psychologists to ensure efficacy.
  6. Eliminate financial incentives. Technology companies as well, as advertising networks more generally, must devise ways to prevent purveyors of dis-information from gaining financially.
  7. Crack down on computational amplification. Take stronger and quicker action against automated accounts used to boost content.
  8. Adequately moderate non-English content. Social networks need to invest in technology and staff to monitor mis-, dis- and mal-information in all languages.
  9. Pay attention to audio/visual forms of mis- and dis-information. The problematic term ‘fake news’ has led to an unwarranted fixation on text-based mis- and dis- information. However, our research suggests that fabricated, manipulated or falsely-contextualized visuals are more pervasive than textual falsehoods. We also expect fabricated audio to become an increasing problem. Technology companies must address these formats as well as text.
  10. Provide metadata to trusted partners. The practice of stripping metadata from images and video (for example, location information, capture date and timestamps), although protective of privacy and conservative of data, often complicates verification. Thus, we recommend that trusted partners be provided increased access to such metadata.
  11. Build fact-checking and verification tools. We recommend that technology companies build tools to support the public in fact-checking and verifying rumors and visual content, especially on mobile phones.
  12. Build ‘authenticity engines’. As audio-visual fabrications become more sophisticated, we need the search engines to build out ‘authenticity’ engines and water-marking technologies to provide mechanisms for original material to be surfaced and trusted.
  13. Work on solutions specifically aimed at minimising the impact of filter bubbles:
    • Let users customize feed and search algorithms. Users should be given the chance to consciously change the algorithms that populate their social feeds and search results. For example, they should be able to choose to see diverse political content or a greater amount of international content in their social feeds.
    • Diversify exposure to different people and views. The existing algorithmic technology on the social networks that provides suggestions for pages, accounts, or topics to follow should be designed to provide exposure to different types of content and people. There should be a clear indication that this is being surfaced deliberately. While the views or content might be uncomfortable or challenging, it is necessary to have an awareness of different perspectives.
    • Allow users to consume information privately. To minimize performative influences on information consumption, we recommend that technology companies provide more options for users to consume content privately, instead of publicizing everything they ‘like’ or ‘follow’.
    • Change the terminology used by the social networks. Three common concepts of the social platforms unconsciously affect how we avoid different views and remain in our echo chambers. ‘To follow’, for most people, subconsciously implies a kind of agreement, so it emotionally creates a resistance against exposure to diverse opinion. ‘Friend’ also connotes a type of bond you wouldn’t want to have with those you strongly disagree with, but are curious about. So it is also the case with ‘like’, when you want to start reading a certain publication on Facebook. We should instead institute neutral labels such as connecting to someone, subscribing to a publication, bookmarking a story, etc.

What could national governments do?

  1. Commission research to map information disorder. National governments should commission research studies to examine information disorder within their respective countries, using the conceptual map provided in this report. What types of information disorder are most common? Which platforms are the primary vehicles for dissemination? What research has been carried out that examines audience responses to this type of content in specific countries? The methodology should be consistent across these research studies, so that different countries can be accurately compared.
  2. Regulate ad networks. While the platforms are taking steps to prevent fabricated ‘news’ sites from making money, other networks are stepping in to fill the gap. States should draft regulations to prevent any advertising from appearing on these sites.
  3. Require transparency around Facebook ads. There is currently no oversight in terms of who purchases ads on Facebook, what ads they purchase and which users are targeted. National governments should demand transparency about these ads so that ad purchasers and Facebook can be held accountable.
  4. Support public service media organisations and local news outlets. The financial strains placed on news organisations in recent years has led to ‘news deserts’ in certain areas. If we are serious about reducing the impact of information disorder, supporting quality journalism initiatives at the local, regional and national level needs to be a priority.
  5. Roll out advanced cyber-security training. Many government institutions use bespoke computer systems that are incredibly easy to hack, enabling the theft of data and the generation of mal-information. Training should be available at all levels of government to ensure everyone understands digital security best practices, and to prevent attempts at hacking and phishing.
  6. Enforce minimum levels of public service news on to the platforms. Encourage platforms to work with independent public media organisations to integrate quality news and analysis into users’ feeds.

What could media organisations do?

  1. Collaborate. It makes little sense to have journalists at different news organisations fact-checking the same claims or debunking the same visual content. When it comes to debunking mis- or dis-information, there should be no ‘scoop’ or ‘exclusive’. Thus, we argue that newsrooms and fact-checking organisations should collaborate to prevent duplications of effort and free journalists to focus on other investigations.
  2. Agree on policies of strategic silence. News organisations should work on best practices for avoiding being manipulated by those who want to amplify mal- or dis-information.
  3. Ensure strong ethical standards across all media. News organizations have been known to sensationalize headlines on Facebook in ways that wouldn’t be accepted on their own websites. News organizations should enforce the same content standards, irrespective of where their content is placed.
  4. Debunk sources as well as content. News organisations are getting better at fact- checking and debunking rumours and visual content, but they must also learn to track the sources behind a piece of content in real time. When content is being pushed out by bot networks, or loosely organised groups of people with an agenda, news organisations should be identifying this as quickly as possible. This will require journalists to have computer programming expertise.
  5. Produce more segments and features about critical information consumption. The news media should produce more segments and features which teach audiences how to be critical of content they consume. When they write debunks, they should explain to the audience how the process of verification was undertaken.
  6. Tell stories about the scale and threat posed by information disorder. News and media organisations have a responsibility to educate audiences about the scale of information pollution worldwide and the implications society faces because of it, in terms of undermining trust in institutions, threatening democratic principles and inflaming divisions based on nationalism, religion, ethnicity, race, class, sexuality or gender.
  7. Focus on improving the quality of headlines. User behaviour shows the patterns by which people skim headlines on social networks without clicking through to the whole article. It therefore places greater responsibility on news outlets to write headlines with care. Research using natural language processing techniques are starting to automatically assess whether headlines are overstating the evidence available in the text of the article. This might prevent some of the more irresponsible headlines from appearing.
  8. Don’t disseminate fabricated content. News organisations need to improve standards around publishing and broadcasting information and content sourced from the social web. There is also a responsibility to ensure appropriate use of headlines, visuals, captions and statistics in news output. Clickbait headlines, the misleading use of statistics and unattributed quotes are all adding to the polluted information ecosystem.

What could civil society do?

  1. Educate the public about the threat of information disorder. There is a need to educate people about the persuasive techniques that are used by those spreading dis- and mal-information, as well as a need to educate people about the risks of information disorder to society (i.e., sowing distrust in official sources and dividing political parties, religions, races and classes).
  2. Act as honest brokers. Non-profits and independent groups can act as honest brokers, bringing together different players in the fight against information disorder, including technology companies, newsrooms, research institutes, policy-makers, politicians and governments.

What could education ministries do?

  1. Work internationally to create a standardized news literacy curriculum. Such a curriculum should be for all ages, be based on best practices and focus on adaptable research skills, critical assessment of information sources, the influence of emotion on critical thinking and the inner workings and implications of algorithms and artificial intelligence.
  2. Work with libraries. Libraries are one of the few institutions where trust has not declined, and, for people no longer in full time education, they are a critical resource for teaching the skills required for navigating the digital ecosystem. We must ensure communities can access both online and offline news and digital literacy materials via their local libraries.
  3. Update journalism school curricula. Ensure journalism schools teach computational monitoring and forensic verification techniques for finding and authenticating content circulating on the social web, as well as best practices for reporting on information disorder.

What could Grant-Making Foundations do?

  1. Provide support for testing solutions. In this rush for solutions, it is tempting to support initiatives that ‘seem’ appropriate. We need to ensure there is sufficient money to support the testing of any solutions. For example, with news literacy projects, we need to ensure money is being spent to assess what types of materials and teaching methodology are having the most impact. It is vital that academics are connecting with practitioners working in many different industries as solutions are designed and tested. Rather than small grants to multiple stakeholders, we need fewer, bigger grants for ambitious multi-partner, international research groups and initiatives.
  2. Support technological solutions. While the technology companies should be required to build out a number of solutions themselves, providing funding for smaller startups to design, test and innovate in this space is crucial. Many solutions need to be rolled out across the social platforms and search engines. These should not be developed as proprietary technology.
  3. Support programs teaching people critical research and information skills. We must provide financial support for journalistic initiatives which attempt to help audiences navigate their information ecosystems, such as public-service media, local news media and educators teaching fact-checking and verification skills.

About the report | Download the report

Leave a Reply

Your email address will not be published. Required fields are marked *

A project of the Harvard Kennedy School’s Shorenstein Center, First Draft News is an open-access site that provides practical and ethical guidance in how to find, verify and publish content sourced from the social web.