Instagram DMs Are Now as Toxic as the Rest of the Internet
New research shows women are being bombarded with abuse. And 90% of the time Instagram does nothing about it.
Many Internet truths are self-evident. Comments are written mainly by trolls, Twitter is filled with hate-generating bots, and Facebook is a global marketplace for disinformation sold to the highest bidder. But research shows a new online vector for bile and misogyny that is much less public: the Instagram DM.
The CCDH (Center for Countering Digital Hate) has studied how social networks moderate, or fail to moderate, a variety of content, including Covid-19 misinformation, identity-based hate, and climate denial. New research shows:
Instagram failed to act on 90% of abuse sent via DM to women included in the study.
Further, Instagram failed to act on 9 in 10 violent threats over DM reported using its tools and failed to act on any image-based sexual abuse within 48 hours.
This study found some of the laxest enforcement of terms of service conditions. Indeed, one of every 15 messages researchers looked at broke the company's rules on abuse and harassment.
To someone like me, one in fifteen may not seem like a lot. As a white, middle-aged CIS male with a modest public profile, I barely get a DM a day. For many women, Instagram is a very different experience.
The CCDH worked with five women with a total of 4.8M followers to create a sample. They included:
Amber Heard, actress, and UN Human Rights Champion
Rachel Riley, broadcaster and CCDH Ambassador
Jamie Klingler, co-founder of Reclaim These Streets
Bryony Gordon, award-winning journalist, and mental health campaigner
Sharan Dhaliwal, founder of South Asian culture magazine Burnt Roti
Each agreed to hand over their message history to the CCDH for analysis. Although Heard and Gordon could not obtain their full data downloads, it serves as a decent sample of women with public profiles that are not necessarily mega-celebrities.
The results are not pretty.
Source: CCDH. 2022
The fundamental problem is that DM requests are largely unvetted by Instagram. That means the first message can be abusive or pornographic, and the user doesn't know until opening it. But even after reporting a message, the CCDH found that Instagram rarely banned or blocked the user.
Instagram Fails to Act on 90% of Abusive Content
Source: CCDH. 2022
In total, participants were sent 482 abusive images and video DMs from strangers, of which 125 fit the definition of image-based sexual abuse, including:
66 examples of users sending women pornography without consent
50 examples of men sharing images or videos of their genitals
Nine examples of pornography edited to feature other faces, known as “fake porn”
Not only is this gross, but it is also illegal in much of the world. Germany, France, Ireland, and Canada all ban so-called “cyber flashing.” Texas is the only state in the U.S. to have similar protections, but similar legislation is moving through other state legislatures.
The #MeToo movement revealed that women experience the workplace very differently than men. Reports like this one show that the divide is just as deep online. The good news is that social media platforms have technologies for moderation that can address these issues.
The same tools that are used to identify your shopping preferences can be used to detect and remove offensive content. If Meta can identify you across multiple sites and know you are in the market for a pair of new sneakers, they can track serial abusers as well.
Last year, Meta announced new, stronger harassment policies: “Public figures shouldn’t be subjected to degrading or sexualized attacks.” Antigone Davis, Global Head of Safety, wrote. Davis said the company already removes attacks on public figures that encompass a wide range of harms but would go a step further. Davis promised the company would also remove:
Severe sexualizing content
Profiles, Pages, Groups, or Events dedicated to sexualizing the public figure
Derogatory, sexualized photoshopped images and drawings
Attacks through negative physical descriptions that are tagged to mention or post on the public figure’s account
Degrading content depicting individuals in the process of bodily functions
Sounds like a good start, but it seems like Meta, the parent company of Instagram, overlooked the DMs.
The CCDH is a registered non-profit. You can support their work here.
Today’s Bits
What Happened At Techonomy Climate
David Kirkpatrick’s excellent wrap-up of the event.
Mark Zuckerberg’s vision for the metaverse is off to an abysmal start
To be fair, it is very early, but there are some solid points here.
Google Says AI Generated Content Is Against Guidelines
I suspect they will reverse this, but it is good, short-term news for human word mechanics.
How to Become a Master Automator
Automate or be automated. The good news is that you don’t have to learn to code to do it.
Buccaneers' Tom Brady Buys Bored Ape Yacht Club NFT for over $430K
Dude, you make it really hard to be a fan sometimes.
Meta Machined
If this is the 12th newsletter, that means I have been doing this for three weeks. To be honest, it feels longer. That is a good thing. I can feel and see the progress we are making. After the initial rush, we are on the long, slow upward climb of attracting subscribers. Suffice to say, I miss my AudDev team at PCMag. Nonetheless, I want to thank all of you for taking this ride with me. If you like the newsletter, please consider sharing it on your socials.
Machined Podcast
I know I promised to kick off the podcast in April, but I landed a big project that might delay that a bit. Rest assured, I am working on it. If you’d like to be a guest on the podcast, hit me up at dc@dancosta.com.