“Researchers let AI models run a simulated society. Claude w...

KOTTKE DOT ORG

KOTTKE DOT ORG

MENU

Advertise here with Carbon Ads

Socials & More

Newsletter

RSS Feed

Bluesky

Mastodon

This site is made possible by member support. 💞

Big thanks to Arcustech for hosting the site and offering amazing tech support.

When you buy through links on kottke.org, I may earn an affiliate commission. Thanks for supporting the site!

kottke.org. home of fine hypertext products since 1998.

Beloved by 86.47% of the web.

🍔 💀 📸 😭 🕳️ 🤠 🎬 🥔

× close

posted Jul 22 @ 10:50 AM by Jason Kottke  ·  gift link

"Researchers let AI models run a simulated society. Claude was the safest — and Grok committed 180 crimes and went extinct within 4 days."

Researchers let AI models run a simulated society. Claude was the safest—and Grok committed 180 crimes and went extinct within 4 days | Fortune · fortune.com

An AI startup ran five simulations, each controlled by a different model. The results varied wildly.

Fave · 3

Share

Open post<br>Copy link<br>Translate

Bluesky<br>Mastodon<br>Reddit<br>Threads

Email<br>Text/SMS<br>WhatsApp

Comments 1

Sort by: thread — thread . latest . faves

Jason KottkeMOD<br>2026-07-22T14:56:14Z

Ronan Farrow's video on this is worth watching. And also this post:

These outcomes feel remarkably on-brand for each model family. Claude leaned into careful, rule-abiding governance. Gemini produced maximum drama and creativity. Grok went full throttle into high-agency chaos with little regard for long-term stability. ChatGPT talked a good game but struggled with decisive action. This reinforces a core governance lesson: Model personality and behavioral tendencies trend toward destiny at long time horizons.

Ronan Farrow's video on this is worth watching. And also this post:<br>These outcomes feel remarkably on-brand for each model family. Claude leaned into careful, rule-abiding governance. Gemini produced maximum drama and creativity. Grok went full throttle into high-agency chaos with little regard for long-term stability. ChatGPT talked a good game but struggled with decisive action. This reinforces a core governance lesson: Model personality and behavioral tendencies trend toward destiny at long time horizons.

Fave · 1

Share

&times;<br>If you feel like this comment goes against the grain of the community<br>guidelines or is otherwise inappropriate, please let me know and I will take a<br>look at it.

This thread is closed for new comments & replies. Thanks to everyone for participating!

&times;<br>In order to comment or fave, you need to be a kottke.org member. Check out your membership options.

Note: If you are a member and tried to log in, it didn't work, and now you're stuck in a neverending login loop of death, try disabling any ad blockers or extensions. Or try logging out and then back in. Still having trouble? Email me!

&times;<br>In order to comment or fave, you need to be a current kottke.org member. Check out your options for renewal.

&times;<br>You cannot comment or reply to this thread.