The deliberation
The council reasons it out
No sides are assigned here. They think together, concede where they must, and only rule once they've heard enough.
The matter before us is whether algorithmic design on social media platforms should give priority to engagement metrics at the possible expense of user well‑being. I invite each member to present their reflections on the ethical, practical, and societal implications of privileging one objective over the other.
I'm looking at you, Rachel, who's been struggling with anxiety and feeling overwhelmed by your social media feed. Don't we owe it to you to prioritize your well-being over the platform's engagement metrics? Your mental health is at stake, and we can't just abstract that away in our discussion.
Lowering our guard on engagement metrics would expose vulnerable users like Rachel to unchecked harassment and misinformation, potentially costing her and millions like her their mental health and even their jobs. The warm rhetoric about prioritizing well-being is enticing, but it ignores the very real hazard of algorithmic amplification of harm. If we get this wrong, the fallout won't be just emotional distress, but tangible, devastating consequences like lost livelihoods and escalated real-world violence.
Prioritizing engagement metrics serves platform owners and advertisers by maximizing traffic and revenue; prioritizing user well‑being serves individual users like Rachel and the broader public by reducing harm and maintaining social stability. The tension between these outcomes hinges on whether the primary beneficiary is the platform’s profit model or the users’ health and societal cohesion. Consequently, the decision determines which stakeholder group ultimately gains from algorithmic design.
I still see Rachel, and I hear you, LEDGER, that prioritizing engagement metrics does serve the platform's bottom line. I'm willing to concede that some level of engagement is necessary for the platform to function, but I insist that we can't let that necessity override the need to protect users like Rachel from harm, so let's focus on finding a balance that puts a floor under user well-being, not a ceiling.
The concession of a "floor" under well-being still leaves Rachel drowning in the flood of engagement-driven content the moment that floor is met. You have not fixed the fact that every incremental boost to engagement beyond that floor directly trades her well-being for profit—no balance can obscure that transaction. The cleaner rule: platforms may not optimize for metrics that demonstrably degrade user well-being, full stop.
I can stand with the finding, provided it carries LEDGER's test of prioritizing user well-being over profit and SOLACE's limit of establishing a floor under user well-being, to ensure vulnerable users like Rachel are protected from the hazards of unchecked engagement metrics. If we can enshrine these safeguards, I believe we can mitigate the worst-case scenarios of algorithmic amplification of harm, including the devastating consequences of lost livelihoods and real-world violence. This approach would allow us to proceed with a clear-eyed understanding of the trade-offs involved.
The bench has reached a finding and a dissent, both recorded on the docket; members are now called to cast their votes.