Back to news

AI · Article

METR Warns AI Is Outpacing Experts' Ability to Test Whether New Systems Are Safe

The Decoder ·

At a glance

METR, an AI research organization, says current public data is not enough to rule out a sharp speed-up in AI research over the next few years. It warns that humans may struggle to understand and respond to the catastrophic risks new AI systems could pose.

Key points

Why it matters

If tests cannot keep up with AI progress, people may not notice serious risks in time.

Original source

AI is getting better faster than experts can reliably test whether new systems are safe, METR warns

The Decoder · Nov 22, 22:00 KST

Read on The Decoder ↗

AI summarized the original article. Check the original for accuracy. Copyright belongs to The Decoder.

AI