SoundCloud Blocks Millions of Bot Actions Weekly, AI Makes Detection Harder

Product Lead Addresses Bot Crisis Ahead of Reddit AMA

SoundCloud Principal Product Manager Harry Manion has delivered the platform's most candid public assessment of its bot problem, revealing that the company now detects and blocks millions of inauthentic actions every week. Manion posted the comments in the platform's verified subreddit ahead of a scheduled AMA session on September 14th.

The product lead acknowledged that artificial intelligence has fundamentally shifted the battle against automated accounts, helping bots defeat systems that were previously effective at filtering fake engagement. He described the dynamic as a continuous game of cat and mouse, one that requires constant vigilance as AI makes it harder to distinguish bot behavior from genuine user activity.

Inside the Bot Ecosystem

Manion said he has invested significant time studying the problem, including going undercover in Telegram groups operated by SoundCloud bot makers. He framed the impact in stark terms, writing that every 1 bot experience devalues the next 10 real ones you get.

While he declined to share tactical details about the company's anti-bot roadmap, Manion outlined three strategic priorities: making bot accounts harder to create and maintain, building better trust signals from user behavior, and giving users simpler tools to filter and flag spam that slips through.

The Human Side of Platform Health

Manion also placed responsibility on legitimate users to help the platform feel more authentic. He noted that many genuine accounts exist as empty profiles, which can make their interactions feel spammy even when they're real. Simple actions like adding a profile picture, writing a bio, leaving comments, and messaging artists to express appreciation all contribute to a more human and connected environment, he said.

The full AMA is scheduled to begin on September 14th at 2pm ET (11am PT) in SoundCloud's verified subreddit.