Michael West Media Builds AI Bot to Hide Hate

The outlet's comment tool reads every post across its social accounts and hides threats and abuse while leaving criticism visible.

Updated on Nov. 19, 2026 in Internet

Flat-color editorial illustration of small tokens in cream and navy flowing along a channel, with navy tokens held behind a grid barrier while cream tokens pass through.
Michael West Media has built an AI bot that reads all comments across its seven social accounts and hides threats and abuse while leaving criticism visible. AI Illustration. Upload story photo >

Live Poll

Do you think online hate speech has gotten worse over the past several years?

Michael West Media has built an AI comment bot that hides threatening and abusive comments across its Facebook, Instagram, X, Bluesky, YouTube, LinkedIn and TikTok accounts. The outlet receives about 500,000 comments a year and says the bot hides roughly one in 80 of them.

Why it matters

The outlet reports that hateful comments on its accounts rose from two in 1,000 in 2018 to about one in 25, a rise that has prompted a technical response to moderation. Criticism of governments, parties and the outlet remains visible under the bot's rules.

The figures come from the outlet's own moderation records, showing hateful comments rising from two in 1,000 in 2018 to about one in 25 now. The source does not report an independent audit of how accurately the bot classifies comments.

The players

Michael West Media

Michael West Media is an Australian independent news outlet that publishes investigative and political reporting online.

The details

The bot reads every comment across the outlet's Facebook, Instagram, X, Bluesky, YouTube, LinkedIn and TikTok accounts and classifies each one. Comments flagged as threats, calls for violence, or abuse targeting race, religion, gender or sexuality are hidden rather than deleted.

Timeline

  1. 2018: Hateful comments stood at two in 1,000 on the outlet's accounts.

  2. 19 November 2026: Michael West Media published its account of the comment bot and the hate figures.

The Tech Race

Publishers have increasingly turned to automated classifiers to manage comment sections as platform moderation has changed. Michael West Media's bot applies that approach to its own accounts across seven social networks.

Readers who comment on Michael West Media's posts may find abusive or threatening comments hidden while criticism of governments, parties and the outlet stays visible. The rule separates hostile abuse from political critique in practice.

The takeaway

Automated moderation can reduce visible abuse at scale, but its usefulness depends on how clearly the rules separate abuse from legitimate criticism. Readers judging any such tool should ask how its classifications are checked.

Further reading

For more on how online platforms handle abuse, see the Internet section.

Source note: This article includes information reported by Michael West.

Live Poll

Do you think online hate speech has gotten worse over the past several years?