06 February 2024
In 2018 I attended a conference in San Francisco by the World Economic Forum. The topic was the creation of “quality signals” in news media – signals which could be attached to a news outlet or deliverable indicating its trustworthiness, and help beat fake news.
The conference was held in a WEF building called the Centre for the Fourth Industrial Revolution, nestled in a beautiful parkland setting called El Presidio, just south of the famous Golden Gate Bridge.
The first day opened with breakfast, where a colleague and I got into a conversation with a professor about the fundamental nature of truth. I gave the example of the famous bridge, which was, I contended, a big red bridge. We could all see it, I suggested, and should be able to reasonably agree upon its existence. But, the professor countered, we had no objective knowledge of the bridge, only subjective experiences of perception of the bridge.
Yes, I wanted to respond, but in your subjective experience, my colleague’s, and mine, the bridge is a bridge, the bridge is red, and the bridge is big. Three out of three, three times. What’s that?, I wanted to ask, a coincidence?
But then the bell rang, and formal proceedings began. It was downhill from there.
The key figures in bringing the conference together, as far as I can tell, were; Mike Hanley who was then Head of Digital Communications for the WEF (he now does something something content for the United Nations) Jeff Jarvis, a professor of journalism at CUNY and the conference’s intellectual centre of gravity, and Craig Newmark, billionaire founder of Craigslist, who paid for the drinks.
In lists we trust?
According to my preparatory notes – the biggest chunk of attendees, 11 of the total 38 – were from academic institutions. In an equal tie for second place, with six each, were representatives of established blue chip tech platforms (Facebook, Google, etc) and a category I called “list/label/club”.
These initiatives work thusly; A group of publications and/or NGOs and/or academic departments band together, attempting to pool their credibility, and call themselves The Trust Project, or the Journalism Trust Initiative, or the Credibility Coalition, or the Certified Content Coalition. The Credibility Coalition has a Credibility Catalogue, by the way, which includes the Certified Content Coalition. This is, I am sure, a huge relief to read. I mean, just imagine if the Certified Content Coalition weren’t included in the Credibility Coalition’s Credibility Catalogue?
These initiatives have different ways of producing their lists, but there are some important commonalities. They include:
- The lists are of institutions, not individual journalists, nor individual stories – meaning they boil down, essentially, to a whitelist of approved URLs, which get some kind of tick of approval (aka a “quality signal”).
- The role of the general public in these schemes is to see the signal and accept that since the outlet has been approved by some unknown set of well paid professionals, they are trustworthy.
The intellectual bankruptcy of such an approach seems to me self-evident. It seems equally evident that such an approach is doomed to make the issue of distrust worse by creating the perception of undeserving media elites condescending to the general public.
An example of abject failure
The Poynter Institute, which is also linked to Mr Newmark’s philanthropic endeavours, was also in the mix. They run something called the International Fact Checking Network, which has a list of certified fact checkers – upon whom Facebook has agreed to depend.
When Mark Zuckerberg went before congress to testify about these issues he faced harsh questioning from congresswomen, Alexandria Ocasio-Cortez, who asked why The Daily Caller, a right wing outlet which she alleged has “ties to white supremacists”, was included. Zuckerberg answered that they had been chosen by the “Independent Fact Checking Network (sic)”. He didn’t get the name right but he was sure they had “rigorous standards”.
Building something new
Stone Transparency, the company I was representing and still operate, was one of only three really novel tools present at the conference. The other two were;
- Factmata, represented by it’s CEO Dhruv Gulati, which at the time was attempting to detect disinformation and hate speech with AI, but which seems to have pivoted to using it to identify “the narratives and influencers behind new brand risks”, whatever shit that means, according to its very broken looking website.
- Hypothes.is, which allows users to annotate pdfs and websites – so that others can see their annotations, was represented by its CEO Dan Whaley.
Both approaches, like Stone, focus on an individual story, rather than slapping a label on an entire outlet. Hypothesis even enlists the public as allies, with a faith that the annotations of one reader will be useful to subsequent readers. That’s an improvement on the listmakers. But there’s something that they have in common with them.
Their system comes into play only after the journalist has published. Everything happens the same up to that point, then, a trust label, or an annotation, or an algorithm, is added, after everything important has already happened.
The journalism process itself remains a walled garden.
Transparency of process and provenance
While some build walls, others are building windows.
The principle is simple – a news deliverable should include, or be accompanied by, details about the work behind it, as an answer to the twin questions “how do you know this?” and “why should I trust you”.
This was the thinking behind the (now aborted) Open Notebook project, run by Hearken – a “a social impact consultancy” from the United States (not to be confused with opennotebook.com – which does “story behind the story interviews” as part of reporting generally on the industry and craft of science journalism).
“The goal”, they wrote in 2017 “is to bake transparency into a reporter’s normal workflow”, this is, in principle, the right idea. They had the right what, and even the right why, but they fell down on the how.
Their own words it was “a mix of mini-newsletter and live blog about a single story”. This hodgepodge of good but underdeveloped ideas did not work, and the opennotebook.co website is now offline.
The problem is not bias, it’s quality
The execution was fumbled, but the idea is sound: getting journalists to show their working. We all agree that research is valuable, so let’s make it visible. For a journalist to be willing to engage in this process, inviting unprecedented accountability is itself a “quality signal”. And if it is handled correctly, it can be shared in a way that can be interrogated by the reader.
Most journalists won’t like it. Most content on the internet is a mindless repetition, copy pasted and reworked from a wire service or a press release without any independent verification, a process which is more likely to introduce an error than add value.
There are of course exceptions, but they are few and far between. Most “journalism” is little more than copying and pasting, interspersed with twitter doom-scrolling.Journalists chase trends and rush to publish spending essentially no time doing original research. The ecosystem favours free-riders. The problem with journalism is not bias – it’s quality.
Body cams for journalists
The crucial difference with Stone is that our tool captures the research process behind a story as a new species of video content. As the journalist works on a story, it captures the content of their screen, and allows them to add webcam commentaries, called highlights, at key moments, which are then embedded in the article, like so:
That’s a research portal, it contains the 17 highlights I recorded during the 231 minutes I spent researching this article. If you click “see research” it will take you to the project, where you can also see the non-highlighted parts, captured as a timelapse – one screenshot per second.
It also means the easy addition of relevant video content to any well researched story – video, as we like to say, without the pivot.
Essentially, we want Stone to be to journalists what body-cams are to cops – a tool for transparency and accountability. Unsurprisingly it’s been an uphill battle.
This approach precludes participation by those producing false or low quality news content. The Stone user community, which already includes some incredible investigative and scientific journalists, can form the kernel of a new, intellectually sound, civic discourse. First this will be a novelty – an elite group holding itself to higher standards – and reaping the rewards including a 45% increase in donations, according to our initial data. Then as more people join, the question increasingly becomes why wouldn’t you use it? What have you got to hide?
Thus we have a glide path from where we are now – with a tiny number of elite early users – to a world where transparency is the new normal, and the disinformation crisis is a thing of the past.