Written evidence submitted by Disinformation Index
The Global Disinformation Index Evidence for DCMS Select Committee on Fake News and the Media
Submitted by Clare Melford and Alexandra Mousavizadeh of Global Disinsformation Index.
March 2018
Executive Summary:
The information ecosystem has become polluted. This has led to a collapse in trust in online news. What can’t be measured, can’t be managed. The Global Disinformation Index (GDI) www.disinformationindex.com is the first global tool for rating the world’s media on the risk of disinformation. It brings together a unique coalition of experts from the London School of Economics; First Draft, Harvard Kennedy School; Oxford Internet Institute; as well as Artificial Intelligence experts from Signal Media (UK) and Veracity.ai (USA). The index will create the most advanced system in the world to rank and rate outlets in real time. It is a UK based not for profit organization.
Due to the many conflicts of interest held by the digital media platforms (Facebook, Google, Twitter etc) we would urge the committee to consider mandating that the companies:
a) make more data available for third party analysis and
b) use neutral, independent external assessors to monitor the extent of disinformation on their platforms
- Definition:
We use the definition of Disinformation laid out by Claire Wardle of First Draft at Harvard.
“Deliberately fake content designed to deceive for financial or political gain”. This excludes bias, satire, malinformation (e.g. correct information distorted or misrepresented to cause harm) and misinformation (e.g. honest mistake).
2. How did we get here:
While political disinformation has been around for hundreds of years, the financially motivated kind has grown hugely in the last decade. There is some evidence that what is now called Information Operations by governments (or political disinformation) has increased sharply in recent years (e.g. Russian IRA activity on Facebook around 2016 US presidential election) presumably as changes in the internet have made it easier and more efficient to deploy.
In addition to an increase in the amount of politically motivated disinformation, two parallel changes in the online advertising market have lead to:
b) huge increase in financially motivated disinformation and
c) microtargeting of all (dis)information.
The first change is the rise of programmatic ad buying networks which automatically place adverts next to content that delivers the requested audience, with no human intervention. The second is the sophisticated micro targeting techniques developed by marketers to segment and profile online audiences with ever greater accuracy. While this was originally intended to make genuine advertising more effective, it has quickly become an invaluable piece in the toolkit of those wishing to convey a political message, true or false.
The internet has enabled the creation of almost limitless content (of all types) and with it, an almost infinite supply of advertising spaces. This greatly reduces prices and makes it possible for almost anyone to earn money from their content via advertising with relative ease.
Advertisers may have cheaper prices but they now struggle to know whether their brand is appearing alongside bonafide content, disinformation, hate speech, or other content unsafe for their brands. The most click-baity content will attract the larger advertising dollars, all without a human getting involved.
Research is ongoing into how much advertising currently ends up next to mal/mis/disinformation. What we do know is that is it increasing. There is no good estimate for the amount of disinformation in any one language nor what proportion is financially or politically motivated.
3. What solutions can help: Bringing Metrics to a polluted information ecosystem:
While Disinformation is hard to measure, our organisation - The Disinformation Index (GDI) - has brought together global experts to develop a new robust system to bring much needed data and metrics to this field, while amplifying the work of other initiatives.
We are working with experts from Harvard (Claire Wardle) and LSE (Anne Applebaum) as well as data analytics firms (Signal Media and Veracity.ai) and index experts Decima to create the first global rating system for the world’s media that will assign a rating to each outlet based on the probability of that outlet carrying disinformation. In much the same way as credit rating agencies rate countries and financial products with AAA for low risk all the way to Junk status for the most risky investments, so the index will do for media outlets.
Measuring the Disinformation on the world’s media sites will support:
- Advertisers to choke off ad funding to disinformation sites
- Quality Media by providing an independent assessment of veracity
- Public by empowering them with greater context
- Policy makers to develop evidence based policy
- Digital Media by providing neutral, independent assessment of Disinformation
The GDI will enable advertisers to steer their ads away from outlets with higher risk of disinformation, thereby choking off funding and reducing the level of financially motivated disinformation. It will do this by making the probability risk assessments real time, and connected a feed from the index directly to programmatic ad-buying networks allowing ads to be placed only on sites with a rating level agreed with the advertiser.
The GDI will allow media outlets to demonstrate an external benchmark of their veracity to reassure readers and advertisers. It will not make any statement about bias or stance, only a risk probability of disinformation. Two publications at different ends of the political spectrum can have the same rating if they have the same risk of disinformation.
These ratings will give the public readers of news more context about what they are about to read and enable them to make informed choices about what they share on social media. In turn if social media platforms allow those ratings to be seen when content is shared on their platforms, it will give those readers getting their news from their social media feeds the same context as those reading from primary sources, enabling informed choices about what to share.
From all the ratings of media outlets in a country we can aggregate a Disinformation Probability for a country. Over time this will begin to show us trends in the geographic spread of disinformation around the world, highlight hot spots and identify countries that are doing well in tackling the problem. This should help policy makers make evidence based policy decisions about, for example, targeted development support to a country.
An example of what the index might look like, (with dummy data) is show below.
A probability rating is not a value judgement. It does not censor. It allows people to make their own, informed choices about how much credence to put into what they engage with. So finally, the index if used by Social Media platforms as an independent assessor of content will allow them to steer clear of being arbiters of the truth, a role they do not want and should not have. And by giving people information about a site before they choose to share it on social media platforms, it will reduce the amount of disinformation appearing on those platforms in the first place.
The Global Disinformation Index is an ambitious project that aims to unite many of the great initiatives working on Disinformation by using their outputs as inputs into the index. The GDI can incorporate the output from fact checkers (such as Snopes or Full Fact) on the number of debunks by a site has received. It can incorporate the output from the News Quality Score or the Credibility Indicators. By providing a big tent to unite all the initiatives working to reduce disinformation, we can amplify the work of the whole field, find solutions faster and stay ahead in the arms race against the purveyors of disinformation.
4. Summary:
The Global Disinformation Index will reduce the amount of financially motivated disinformation by enabling advertisers to steer their ads away from high risk media sources.
The index will benefit the public by providing context which may make them less likely to share high risk content. It will benefit policy makers by providing metrics on countries and trends over time, and it will benefit the media companies by providing external mark of risk probability. Finally we think an index will help the digital media platforms monitor and manage the extent of disinformation on their platforms.
We are currently fundraising and, if successful in securing funds, we aim to have the English language prototype complete for 5 countries by Q1 2019.
For more information please visit The Disinformation Index
5: What the Select Committee can do:
We call on the committee to recommend that digital media companies be compelled to make data available for analysis by third parties to speed up the fight against all forms of information pollution.
Secondly we call on the committee to recommend that digital media companies be subject to external assessment of the extent of disinformation on their platforms. We do not make a statement about whether this should be regulation or self regulation, but only that an outside body be tasked with assessing disinformation in the UK and elsewhere, and that the companies are required to allow the assessment and make public the results.
March 2018