Social media takes new approach to online commentary

Facebook’s self-regulatory contortions in the wake of fake news and trolling scandals have gone on, with little visible effect, for months.

Now Twitter founder and chief executive Jack Dorsey has announced his company is going to try a different tack – but Mr Dorsey’s approach is arguably even more far-fetched than his Facebook peer Mark Zuckerberg’s: it’s an attempt to view Twitter’s social mess as an engineering problem.

In a Twitter thread last week, Mr Dorsey admitted that Twitter has been home to “abuse, harassment, troll armies, manipulation through bots and human-coordination, misinformation campaigns and increasingly divisive echo chambers” and that it’s not proud of how it has dealt with them. So it would try to find a “holistic” solution through attempting to “measure the ‘health’ of conversation on Twitter”. The metrics, designed in collaboration with outside experts, would presumably help redesign the service so that all the bad stuff would be gone without the need for censorship.

That’s not how Facebook chose to handle a similar problem. For starters, it didn’t ask anyone for advice.

Facebook just devised some possible solutions such as working with fact-checkers to identify fake news and focusing on content from friends rather than publishers; it even experimented with putting publisher content in a separate newsfeed – a test it has just ended because users apparently didn’t want two feeds. It has also volunteered to reveal more information about who bought political ads.

It’s not clear whether these moves have done anything to fix the problems: I still have my tens of thousands of fake “subscribers” who showed up after I was active in the 2011 protests in Moscow and, as far as I’ve seen, questionable content from highly partisan sources is also still there. All that has happened is that, according to a recent analysis of Nielsen data by equity research company Pivotal Research Group, time spent by users on Facebook was down 4 per cent year-on-year in November, 2017.

Twitter, faced with a 14 per cent decline in time spent and a decrease in attention share to 0.8 per cent from 1.1 per cent over the same period – and consequently described by Pivotal Research Group as a “niche platform” – needs to do something that will draw people to it, not repel them. So one can understand software developer Mr Dorsey’s need for a bottom-up re-evaluation of how his software has been working.

The starting point has been provided by a non-profit called Cortico, which grew out of the Massachusetts Institute of Technology Media Lab. It’s working on a set of “health indicators” for the US public sphere based on four principles: shared attention (to what extent people are interested in the same subjects?), shared reality (are people using the same set of facts?), variety (are people exposed to different opinions?) and receptivity (are they willing to listen to those different opinions?).

If there’s a transparently developed, openly discussed set of measurements to determine the “health of the conversation” on Twitter or any other social network, the networks could, instead of grappling with macro-problems like “fake news” or “harassment”, break down their responses into micro-actions designed to move the metrics. Then they could report to the public that the conversation is growing healthier.

It’s a bitter pill for the networks’ engineer founders, who tend to think technology and data can fix any problem, but a solution can only be found by putting people in the same environment that exists in face-to-face conversation or in the “legacy” news media – one that makes outbursts and lying legally and socially costly.