Facebook's 2018 News Feed change offers a concrete dispute about how a system can pursue a stated social aim while the signals used to rank outcomes may create effects that the objective does not capture. This case concerns the design and reported side effects of ranking; it does not claim that engagement alone determines the Feed or that social media's total effect is known.
The 2018 ranking change
In January 2018, Facebook said News Feed would give more weight to posts predicted to produce conversation and meaningful interaction. Its announcement named reactions, comments and sharing among ranking signals, and described a shift toward “meaningful social interactions.” See Facebook's own account: Bringing People Closer Together. On the Q4 2017 earnings call, the company also discussed an expected tradeoff in time spent and some engagement measures (transcript PDF).
Criticism and internal documents
In 2021, Frances Haugen told a U.S. Senate subcommittee that engagement-based ranking and the meaningful-social-interactions change could reward outrage and sensationalism. That claim is her testimony as a former employee, relying on internal materials she supplied; it is not a congressional finding or an independently reproduced causal estimate. See the Senate hearing page and the published hearing record (PDF). The Wall Street Journal reported on internal documents about the MSI change; use that reporting for what its journalists found, and the hearing record for what Haugen testified. An SEC complaint is not an SEC finding.
Company response and ranking complexity
Meta has said its systems were not designed to reward provocative content, described demotion of sensational or misleading content, and explained that ranking combines many predictions and survey-based measures of value — engagement signals and predicted interactions among them, not as a sole objective. See Facebook's April 2021 note on incorporating more feedback into News Feed ranking and Meta's November 2021 response on safe experiences and the business.
What the Framework asks us to notice
The dispute does not require us to settle the total effect of Facebook before seeing the moral problem. Success against a chosen proxy cannot, by itself, answer whether a system serves the people whose attention and relationships it shapes. Answerability begins by making reported side effects visible, testing whether the chosen signals serve flourishing and changing the system when its consequences contradict its stated purpose.
Limits
These sources do not establish that every high-engagement post is harmful, that engagement is Facebook's only ranking aim, or that one company proves the Christian hypothesis. They document a concrete dispute about a system's objective, its observed consequences and responsibility for people an aggregate measure can leave out.
Back to the argument
See Flourishing beyond self-preservation in the Framework essay.