AI safety
AI safety is the work of preventing AI systems from causing harm — from everyday failures such as dangerous advice or biased decisions, to misuse by bad actors, to potential large-scale risks from very capable future systems.
In one line, for a 12-year-old
AI safety is everything people do to make sure AI helps and doesn't hurt.
An example
Before release, a lab tests whether its new model will help someone plan a cyberattack, and adds safeguards if it would.
Why it matters to people
Canada created the Canadian AI Safety Institute in 2024 to study these risks with international partners. Public interest research needs independent funding, not only company labs.