WeeklyTech #109

Is content moderation stifling public discourse?

Content moderation is difficult work for any social media company. Every day millions of posts and messages are shared on these platforms, most are benign in nature but as with anything there will be abusive, hateful, and sometimes violent content shared or promoted by certain individuals and organizations. Most social media companies expect their users to engage on these platforms within a certain set of rules or community standards. These content policies are often decided upon with careful and studied reflection on the gravity of moderation in order to provide a safe and appropriate place for users. It is an admittedly difficult and thorny ethical issue though because social media has become such a massive and integral part of our diverse society, not to mention the hyper politicization of such issues. 

Over the years, content moderation practices have come under intense scrutiny because of the breadth of the policies themselves as well as their misapplication—or more precisely the inconsistent application—of these rules for online conduct. Earlier this year, The Daily Citizen—the news arm of Focus on the Family—was reportedly locked out of their account due to a post about President Biden’s nomination of Dr. Rachel Levine to serve as assistant secretary of health for the U.S. Department of Health and Human Services (HHS). The Daily Citizen’s tweet was flagged by Twitter for violating its policy on hateful conduct, which includes but not limited to “targeted misgendering or deadnaming of transgender individuals.” This broad policy seems to include using the incorrect pronouns for individuals, using the former name of someone after they transition and identify by another name, or—in the case of The Daily Citizen’s tweet—stating the biological and scientific reality of someone’s sex even if they choose to identify as the opposite sex or derivation thereof.


The Rundown

Study: Radiologists better at detecting breast cancer than AI – Hayden Field | Emerging Tech Brew

In this week’s episode of Overhyped AI: When used to detect breast cancer in mammography screenings, algorithms reportedly fall short in a big way—in nearly all cases, they’re less accurate than a radiologist. That’s according to a systematic review conducted by UK researchers and published in The BMJ, a peer-reviewed medical journal. 

Facebook says its AI mislabeling a video of Black men as “primates” was “unacceptable” – Kim Lyons | The Verge

Facebook is apologizing for an incident where its AI mislabeled a video of Black men with a “primates” label, calling it an “unacceptable error” that it was examining to prevent it from happening again.

Tripwire CEO out after tweeting support of Texas abortion law – Stephen Totilo | Axios

The head of Georgia-based game development studio Tripwire Interactive parted ways with the company Monday, two days after tweeting his support for Texas’ new abortion ban.

Australia’s Top Court Finds Media Companies Liable for Other People’s Facebook Comments – Mike Cherney | Wall Street Journal

Australia’s highest court found that newspapers and television stations that post articles on Facebook Inc.’s platform are liable for other Facebook users’ comments on those posts, a ruling that could prompt traditional publishers to rethink how they engage with social media.

Texas Governor Greg Abbott signs social media “censorship” bill into law – Kathryn Watson | CBS News

The new law requires social media companies with more than 50 million monthly users to disclose their content moderation policies and institute an appeals process. It would also require such social media companies to remove illegal content within 48 hours. 

Software that monitors students during tests perpetuates inequality and violates their privacy – Shea Swauger | MIT Technology Review

The coronavirus pandemic has been a boon for the test proctoring industry. About half a dozen companies in the US claim their software can accurately detect and prevent cheating in online tests. Examity, HonorLock, Proctorio, ProctorU, Respondus and others have rapidly grown since colleges and universities switched to remote classes.