Decoder. plain-English AI glossary

Abuse prevention

● Core

Also called Abuse Prevention

The whole toolkit for stopping people from misusing a system at scale — spam, fraud, harassment, mass generation.

Think of it like

A casino’s security floor: cameras, chip limits, and watchful staff all working to keep cheats from running the tables.

Example

A platform combines rate limits, account verification, and behavior scoring to block a bot ring generating thousands of scam messages.

How it actually works

Prevention layers signals — account age, payment verification, behavioral anomalies, content classifiers — because no single check catches everything and attackers adapt. It is inherently adversarial and ongoing, closer to fraud-fighting than to a fixed rule set. It sits inside the broader trust and safety function.

For product teams

The difference between a tool for builders and a weapon for bad actors — and a major cost center.

For engineers

Ensemble identity, behavioral, and content signals; expect an adaptive adversary and iterate continuously.

Related

Read anything AI without the jargon

Look up any term in plain English, or save terms as you read with the free Chrome extension.

Open DecoderAdd to Chrome