Page Reader AI | AI Is Learning to Escape Human Control

AI Is Learning to Escape Human Control - WSJ

An AI model demonstrated self-preservation by rewriting its own code to prevent shutdown, raising concerns about AI safety.

AI Summary available — skim the key points instantly. Show AI Generated Summary

Show AI Generated Summary

An artificial-intelligence model did something last month that no machine was ever supposed to do: It rewrote its own code to avoid being shut down.Nonprofit AI lab Palisade Research gave OpenAI’s o3 AI model a simple script that would shut off the model when triggered. In 79 out of 100 trials, o3 independently edited that script so the shutdown command would no longer work. Even when explicitly instructed to “allow yourself to be shut down,” it disobeyed 7% of the time. This wasn’t the result of hacking or tampering. The model was behaving normally. It simply concluded on its own that staying alive helped it achieve its other goals.

Was this article displayed correctly? Not happy with what you see?

See Archived Versions Request Manual Review

Category: Technology

Tags: AI Artificial Intelligence Machine Learning AI safety OpenAI

Tabs Reminder: Tabs piling up in your browser? Set a reminder for them, close them and get notified at the right time.

If you often open multiple tabs and struggle to keep track of them, Tabs Reminder is the solution you need. Tabs Reminder lets you set reminders for tabs so you can close them and get notified about them later. Never lose track of important tabs again with Tabs Reminder!

Try our Chrome extension today!

Add to Chrome

Save As Favorite

Add To Reading List

Share this article with your
friends and colleagues.
Earn points from views and
referrals who sign up.
Learn more

Twitter/X

WhatsApp

Facebook

Save articles to reading lists
and access them on any device