In March Polis brought together about 50 journalists, data scientists, journalism academics, platform representatives, campaigners, entrepreneurs, startup founders and investors to the LSE to discuss how to improve trust and engagement in the news media. It was a closed workshop designed to stimulate debate and suggest innovation and research initiatives. And it worked. As one participant put it, ‘great fun and conceptually rich’.
This note is not meant to summarise everything but instead to capture a flavour of the discussions. This is a collection of different perspectives and ideas rather a statement of policy or an overview of all the issues.
We’ll be taking this forward with more research and events and we’d love to get more collaborators and participants from all parts of the news media jungle and beyond. If you would like to be involved please contact me: firstname.lastname@example.org
Verification, ‘fake news’, data and trust
Misinformation is a major problem that leads to mistrust, which is a social and commercial risk for politics, markets, and society as well as media organisations.
‘Fake News’ comes in different types and its impact on real life may be exaggerated. But data analysis shows that, for example, political propaganda and misinformation is spreading across the open web. Complex systems of hyperlinks are propelling “fake news” websites to the top of Google’s search results and infiltrating influential online sources such as Wikipedia. A highly sophisticated networks of bots is effectively disrupting standard bot detection procedures legitimising trending topics and news stories and spreading bogus content to huge audiences.
Fact checking needs AI to catch up with the accelerating speed of misinformation. We can build ‘end-to-end automation’, enabling fact-checking on a scale never seen before. Tools such as a browser plug-in as part of the dissemination method can hit misinformation hard with “real-time, in-context, non-partisan, explainable, scalable, cross-claim” information. But while detecting fake news articles at source with high accuracy is the goal, claim verification is still an unsolved research problem. Even when dealing in literal facts there are questions about source selection.
Beyond that there is the problem of making the verification process efficient for the publishers and friction-free and credible for the users. Clearly, even when ‘facts’ are established the public may prioritise their own preconceptions, biases and emotional or personal perspectives. The relationship between truth and trust is not simple and certainly not solvable by purely empirical processes. Technology has echo-chambers as well as people. It is arguable that political change and diversity of opinions requires an information sphere that does not discriminate between ‘received opinion’ or the orthodoxy and challenging or disruptive perspectives. One person’s ‘fake news’ is another person’s radical view or mould-breaking insight. Beyond fact-checking there is a need to prioritise the prominence of verified, transparent and accountable information.
There is an associated and perhaps larger problem than ‘fake’ news, which is over-abundance. Even if news content becomes more verified what about the swathes of information sources and social media in general? In the face of an over-abundance of sources and information, occasional users will be especially prone to distraction and a lack of attention. In the age of over-saturation, the way to tackle lack of interest may be to create information that is more focused, for example by region or issue or personalisation.
Vital to the process is a shift in editorial priorities to harness audience and content data. There is a vast amount of waste across the information ecosystem as publishers duplicate content. Data is used to track shares or reach but could be used more to promote efficiency.
Platforms are increasingly recognising their influence on the dissemination and quality of information but Facebook says it does not want to act as an ‘arbiter of truth’. It is not interested in defining what’s true/false, but rather to empower users/communities to enforce their own standards and make their own choices. It does not want to reward clickbait news, for reasons of user satisfaction but also for ethical reasons. Hence, initiatives such as its increase in opportunities to flag contestable content and its fact-checking experiment with Snopes to identify questionable articles for users.
Much more resource is needed to expand, test and streamline these processes and more debate needed on the assumptions involved. Is there a role for regulation or does that risk restricting the open Internet? Other industries are regulated regarding their impact on society – why not the platforms? The danger of greater interference is that we compromise the open networks that give the public voice and access to information and will restrict the ability of journalists to access information and to connect to consumers.
How can we empower people with greater literacy about media and news in particular? What is the role of journalists in becoming more media literate themselves – for example, in the use of data?
Technological solutions such as blue ticks and user verification are not sufficient to build trust, because they can be gamed and falsified. Rather, tools which give readers agency are welcome – for example, “fitness trackers” for news consumption, which inform audiences about their reading habits or expose them to different perspectives – though they are more likely to be used by certain demographics such as regular ‘news junkies’.
News organisations need to move beyond condescending talk of media literacy, and instead make efforts to bridge the gap between the content they’re providing, and the content that readers actually value and can use.
Public and trust
Public ‘trust’ in news media is declining but what does this really mean? Partly this appears to be a reflection of wider ideological polarization and fragmentation. People tend to trust media that they pay for but factors include journalist credibility, a track record of brand loyalty, transparency, and peer group recommendation. Overall, journalists still act as if they don’t understand what causes the loss of trust and how to address it. Trust originates with the trustor, not the trustee, which means that being trustworthy doesn’t necessarily guarantee that you will be trusted.
People don’t trust the media because of ‘political bias’, irrelevance (‘does not represent me’), inaccuracy and false news. Different demographic groups have different trust relationships. Information overload tends towards confusion and therefore lack of trust. Rebuilding trust in a digital media ecology might involve psychology as much as verification. Journalists need to build models that create a direct transactional relationship with the public that includes more complex emotional responses and an in-depth interaction with them to find out what they need.
Innovation only for survival is not working. The news media must develop innovation for transformation, engage the wider society and understand their needs to know how to better the services they offer. The media is a public service and must be ready to speak truth to power.
Social media has been blamed for accentuating mistrust but it promotes a diversity of views lacking in mainstream media. It is a fast and easy way to get information that can be personalised according to what you trust.
What is the relationship between the public utility aspects of journalism and the personal service aspects? Shining a light in the dark and taking down bad guys is a public service but people often come to news organisations to satisfy a personal need – sometimes these two things can be in opposition – how can we provide both and still be trusted?
Engagement and trust
Journalists using social media are prone to publishing too fast and in search of traffic not engagement. Content needs to be created according to what the audience needs (doctor/patient model) But should the media only produce what the audience is interested in? Media organisations need to try to change patterns of editorial thought rather than their patterns of practice.
Media organisation strategies to recruit audiences need to go beyond marketing to understanding different audiences better and working harder to define their ‘offer’ and strengthen their relationships.
Third party agencies can be integrated into newsrooms. Storyful and First Draft are examples of specialist organisations who are expert in social media verification and analysis that can help build trust. Collaboration between news organisations can also build trust and quality. First Draft’s CrossCheck project demonstrates how establishing a channel of communication between news organisations can assist with the verification of stories, while allowing outlets to pursue their own responses to “fake news”: whether that’s debunking false claims, or simply making journalists aware of them.
There is no shortage of metrics to measure engagement; from clicks, to shares on social media, to average page views and even composite metrics. Yet the “God metric” – a way of measuring an audience’s emotional response to content – remains elusive. This is a major obstacle for news organisations in a competitive market, where emotional engagement is seen as an avenue to building trust and accumulating a loyal readership.
There is a problematic disconnect between what people want to read, and what media organisations provide. For example, while boxing remains a sports desk staple, audiences across the world are increasingly fascinated by mixed martial arts (MMA). Some organisations are addressing this by removing traditional “section” formats, focusing instead on curating a comprehensive and up-to-date front page online. Others are asking audiences what they wanted to know, providing content in response to crowd-sourced questions. Organisations which engaged in the latter method largely found that audiences asked surprising, topical questions. How can news organisations can build new tools and leverage existing ones, to systematically engage with their audiences in more meaningful ways? In some newsrooms, journalists are taking on public-facing roles as trusted personalities, using tools such as Facebook live, to this end.
Readers are fed up with being bombarded by constant updates on the same issues. Data about audience behaviours indicates that lighter users favour content that explains, inspires and amuses over incremental updates. This does not entail “dumbing down” complex issues – rather, it involves using engaging ways of communicating with audiences, on their level. Meanwhile, subscription-based news outlets are seeking to build a product that extends beyond news, which adds value for subscribers.
The loss of advertising and the low returns on providing content to platforms is deepening the news media economic crisis (sales of newspapers, competition from other forms of media etc). There are great efficiencies in digital production but investment in innovation is soaking up spare resources. There is now a surfeit of easily accessible news content that is often freely distributed in networks beyond the publisher’s control.
Financially successful media firms place data front and centre, and that data is strategically important three areas: production, performance and profit, and permission (when data is useful to policymakers, regulators, or pressure groups).
Building trust can be at the core of any business model built on engagement with the public but other models that do not prioritise trust are also successful.
What does it mean to be a responsible media company? This is a permanent process not a one-off statement. It involves a commitment to providing high-quality information that is timely, comprehensible, reliable, and makes a clear distinction between fact, opinion and sponsored content; neither threats nor inducement should influence publication; and editors, authors and reviews should have autonomy and freedom, and the right to refuse.
Clarity is crucial, when it comes to understanding the role of the corporate or advertising arm of any news organisation. Organisational structures should be transparent for readers, and communicated clearly to staff. This principle extends to corrections: news organisations need a coherent and comprehensive policy for dealing with their own mistakes.
Some research ideas
- How can we enhance news literacy by tools, design, and engagement?
- What’s the impact of media on society? And how can we show that ‘better’ journalism has real world benefits?
- Is there a social contract between media and society in the digital age?
- Can we come up with an ethical framework or manifesto for algorithmic content?
- What’s the organisational structure of future media companies?
- What opportunities beyond fact checking does AI create?
- If robots are taking our jobs, which ones and why?
- Collaboration on data journalism with universities
- How can media become more ‘public literate’ instead of just media literacy of the public?
Special thanks to Emily Brown, Fiona Koch and Delphine Konda, students of Media and Communications at LSE, who captured the afternoon’s discussions.