Channel sheet · CH-03 · gain 3 min · logged November 11, 2026

Social & Influencer AIDirect input

X Previews AI Labels That Explain Why Posts Get Flagged

X previewed AI moderation labels that explain flag decisions and allow appeals, as its new safety chief pitches auditable AI review to win back advertisers.

By Sophie Lindqvist3 min read520 words

Signal notes

  1. X previewed AI-generated moderation labels at the Brand & AI Safety Summit in New York on Nov. 10
  2. The labels explain what the AI detected, show how reach was limited, and allow appeals
  3. X's AI once mislabeled a SpaceX rocket launch post as 'adult content' with no visible reason
  4. X's advertising revenue has plummeted since Elon Musk took over the company
X Previews Labels That Tell Users Why Its AI System Flagged Their Posts
Input monitorX Previews Labels That Tell Users Why Its AI System Flagged Their Posts — AI-generated

X showed off moderation labels on Tuesday (Nov. 10) that would tell users exactly why its AI flagged their posts and let them appeal the decision. Michael O'Herlihy, the platform's new head of safety, previewed the designs at the Brand & AI Safety Summit in New York.

The preview answers a complaint that has plagued X's automated moderation for years. X's AI-powered system once mistakenly labeled a post featuring an image of a SpaceX rocket launch as "adult content." No one, including the author of the post, could see why the decision had been made. Appeals under the current system give users little to work with.

What would the new labels do?

According to the designs O'Herlihy showed at the summit, the labels would appear beneath flagged content and do three things:

  • Explain what the AI detected in a post
  • Detail how exactly the post's reach was limited after the flag
  • Allow the user to appeal the decision directly from the label

O'Herlihy suggested that in the future X's content moderation systems may show their decisioning process, not just the outcome. He is a Twitter veteran who recently rejoined the company as head of safety, so the preview carries the weight of someone who knows the platform's history with automated takedowns.

Why is X pushing AI-led moderation?

O'Herlihy's core argument is that AI-led moderation creates auditable records that work better than human review. If a machine logs every step of its decision, the reasoning goes, that record can be inspected, contested and corrected. Human reviewers, by contrast, rarely leave a trail detailed enough for a user to challenge.

What does this change for advertisers? Potentially a lot. X's advertising revenue has plummeted since Elon Musk took over the company, and brand safety has become a crucial part of X's pitch to win brands back.

Advertisers have historically demanded two things from platforms: consistent enforcement and proof that enforcement works. Transparent labels with an audit trail speak to both. If X can show that its moderation decisions are explainable and reversible, it has a concrete story to sell to media buyers who left over moderation concerns.

What remains unanswered?

The preview is a design exercise, not a shipped product. X did not announce a launch date for the labels, and it did not say whether the appeal process shown in the designs would route to human reviewers or stay fully automated.

There is also the question of accuracy at scale. An AI system that wrongly tags a rocket launch as adult content is one thing; a system that also has to generate a correct, detailed explanation for every flag is a harder engineering problem. The SpaceX mislabel is the cautionary example: the explanation layer only helps if the underlying detection is sound.

For now, the labels remain a signal of direction. X is betting that showing its moderation homework, rather than just issuing verdicts, can rebuild trust with both users and advertisers. Whether that bet lands depends on execution details the company has not yet shared.

via adweek.com (Source)

Filed under

  • x
  • ai-moderation
  • brand-safety
  • content-moderation
  • explainable-ai
Share this article:

More from Sophie Lindqvist

Sophie Lindqvist

Show full bio

Staff writer covering money, funding and market shifts around marketing professionals running AI in the stack at Mart Signal.

164 articles

Bus out

‹ Previous articleNext article ›