Most middle managers aren’t yet equipped to recognize the hidden or long-term costs of relying on AI, including integration overhead, quality-control gaps, and the erosion of early-career talent pipelines.

As part of our research on competing narratives about international migration on social media, we reviewed existing scholarship to develop a typology of 11 frames that capture the dominant ways these debates are structured. These 11 frames underpin the Migration Ads Observatory.

Unfortunately, the social media industry is increasingly moving away from using human fact-checkers. Many are rolling out AI and Community Notes-type programs, which come with their own set of challenges. For example, 89% of Community Notes on X never see daylight. notetracker.socialmediadata.org

Algorithmic audits have become a fashionable remedy in academic and policy debates, but this overlooks a fundamental point: like domestic pets 🐕, every algorithm has an owner. Accountability lies not with the code, but with those who design, deploy, and profit from it.

Like the gun debate in America, where money and identity override reason, the tech industry’s fixation on profit and polarization will make their stated goal of fostering healthy online conversation the Silicon Valley version of “thoughts and prayers.” #emptymantra

In that paper, we showed how an online community legitimizes itself and minimizes allegations of anti-social, bullying behaviour by weaponizing and deploying a feminine gender identity in three overlapping and internally contradictory ways to: socialmedialab.ca/2024/03/12/c...

This new study is a follow-up to a study we released last year, "Just being a bit bitchy: The gendered valences of online anti-social behaviour on Tattle Life". socialmedialab.ca/2024/03/12/c...

PS... To be clear, we are not advocating for toxic speech online, what we are trying to do with this study to understand how toxic speech is used by people to construct and maintain in- and out-group boundaries in online communities.

The Civility Analyzer calculates toxicity scores (such as Toxicity, Insult, & Threat) & prosocial scores (like Compassion, Curiosity, and Respect) for each post.