TikTok may be regretting its appearance before the Oireachtas committee on Arts, Media, Communications, Culture and Sport on Wednesday. The company, which had previously declined to appear, spent an uncomfortable hour or so being questioned by TDs and Senators.
While Meta and Google were also present, much of the ire was directed at TikTok, which, as the platform of choice for young people, has come under scrutiny in recent weeks amid dangerous driving incidents spreading on its platform.
The platform was grilled about the level of content it was seeing, catching and banning before it was recommended to viewers, with Zachary Hecht, TikTok’s head of global trust and safety policy, coming under pointed questioning from the committee, which agreed on one thing: content moderation at the platform isn’t working. TDs and Senators gave numerous examples of videos available on the platforms that breached the platform’s community standards – content that glorified school shooters, misogynistic videos, even some videos of dangerous driving, despite a clampdown.
[ Social media: All you need to knowOpens in new window ]
I am not suggesting that anyone should feel any sympathy here for TikTok, which makes billions of dollars in revenue every year off the back of the millions of videos uploaded each day. It is fair that all the platforms are held to account. In fact, they should come under even more scrutiny, particularly around how reporting systems work.
READ MORE
Have you ever tried to get content removed from social media platforms? Things may have become slightly easier in recent years, as various EU regulations put more obligations on companies to take action against harmful content on their platforms. Trying to categorise your objection is an experience in itself.
If you report content as going against community standards, there is still a chance that your complaint gets rejected because you failed to categorise it correctly.
But there is certain content that seems to prompt a quick (ish) response. Recently, I logged into my Instagram account to a notification that the platform had removed a video I had posted. The topic – Meta’s smart glasses and how easy it was to defeat the “safety” feature that was designed to prevent covert recording – was a newsworthy topic at the time I posted it in August. In September, Meta pushed out an update that stopped people from covering the notification light, rendering the safety feature, such as it is, absolutely useless. Things had moved on.
So it was a surprise to find that not only had the month-old video been removed, but the reason for it was because it may “promote stolen, forged or fake goods and services”. Examples of this offence include selling fake currency, stolen credit cards or promoting services for writing fake reviews. One of the consequences of the content removal for failing to follow guidelines on co-ordinating harm was that I couldn’t go live – something I’ve never done anyway – for 365 days, or monetise content until the policy violations were resolved.
The decision was put at the feet of Instagram’s “technology”. After I requested a review, on the basis that it raised awareness, the video was reinstated. That will appear as a win somewhere in Meta’s regular transparency reports, and probably as a false positive that was subsequently reversed. On further investigation, the mention of “loophole” in the caption potentially triggered an automatic response, and any human moderator missed the context of the video.
But it took a month, possibly due to a renewed push against content that talked people through how to film covertly. A problem that had been repeatedly flagged with Meta weeks before.
So perhaps the committee members’ assessment of the effectiveness of these systems – that they don’t work – wasn’t unfair.
While the company made noises about wanting the platform to be a safe place for its users, the reality of social media is very different. “Harmful” does not always rise to the level of “criminal”.
Controversial content, which can shape people’s opinions and viewpoints, gets elevated because it promotes engagement. So the machine keeps churning out rage bait content that may or may not be removed if the automated systems pick it up or someone manages to get an objection in front of a moderator that understands context and nuance.
Should Ireland share its corporation tax bonanza with the EU?
Let’s put it in context. If TikTok removes 99.6 per cent of videos that it says breach its community standards and rules before they are published, that still leaves enough videos – a little over 1,500 in the first quarter of the year, according to the figures for Ireland in the company’s most recent trust and safety report – that were not caught before they were uploaded. Of the videos removed, 83.4 per cent of those were removed before being viewed, meaning almost 65,000 were viewed. And 88.9 per cent were removed within 24 hours of publication, leaving roughly 43,450 that were not.
What is the solution? More legislation and powers for regulators to act? Media regulator Coimisiún na Méan has six open investigations of its own, alongside participating in European-led inquiries. Despite the regulator’s head Jeremy Godfrey’s assertion that the commission was independent and “fearless”, Sinn Féin’s Joanna Byrne was unimpressed. “Just slow to act, maybe.”
Committee chairman Alan Kelly may think we are at a “watershed moment” in this country. But those of us who have seen similar moments play out in the past suspect that we will still be having these conversations in a year or two’s time, if not more.
