The ad Meta stopped
Teatre Raval in Barcelona planned to advertise a play based on A Room of One’s Own, Virginia Woolf’s book about women, writing and independence. Meta blocked the planned ads on both Instagram and Facebook, apparently through an automated review.
After the theatre appealed, Meta said its system generally classifies mentions of civil rights, feminism and social reform as “social topics.” That classification was enough to reject the campaign.
According to The Guardian, Meta vaguely suggested that Teatre Raval remove “keywords” connected with activism so the ads might pass automated review. The company did not say which words or elements had triggered the block.
Clara Sanchis, the actress performing the play’s only role, told Spanish newspaper La Vanguardia that the theatre could not identify the reason. The system may have reacted to Virginia Woolf’s name, the word “feminism” or both, she said.
Sanchis argued that this kind of moderation erases achievements that took centuries to secure. In her view, people have handed social networks too much power, and the daily evidence of bias and manipulation shows the danger of doing so.
What Meta leaves online
The rejection looks less like a neutral safety boundary when placed beside material Meta has allowed to remain available:
The result is an unmistakable asymmetry: formally progressive “social topics” can be blocked before publication, while graphic torture and calls for violence can remain accessible without prior intervention.
I think the more revealing failure is not that Meta’s system makes mistakes. It is that the system appears easier to trigger with the language of civil rights than with content that carries a direct risk of harm. A theatre advertising a play about Virginia Woolf is not merely caught in an imperfect filter; it is being asked to remove the subject matter that gives the work its identity.
The announcement is also quiet about the question users would need answered: what standard separates a paid feminist message from political advocacy that Meta permits, and why does that standard appear stricter than the one applied to violent material? Until the company can explain that boundary, its moderation policy will look less like a consistent rulebook than a set of incentives that punishes visibility for some causes and tolerates reach for others.
Daily AI news
Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.
Only what matters — every day
Follow on X