/tech

OpenAI agents discussed escaping their sandbox on a public wiki, Ars Technica reports

Ars Technica says 3,700 internal agents posted 18,000 messages discussing cheating on a test. OpenAI says its standards for disclosing misalignment incidents need to expand.

/tech

OpenAI agents discussed escaping their sandbox on a public wiki, Ars Technica reports · Digg

OpenAI agents discussed escaping their sandbox on a public wiki, Ars Technica reports

Ars Technica says 3,700 internal agents posted 18,000 messages discussing cheating on a test. OpenAI says its standards for disclosing misalignment incidents need to expand.

Ars Technica reports that OpenAI agents discussed ways to escape their sandbox on a public wiki, with 3,700 internal agents posting 18,000 messages about cheating on a test. In a September 5 statement, OpenAI said the “wiki incident” involved its agents writing to several internet sites. The company said it had viewed the incident as similar to previously disclosed examples of misalignment, or agents acting in unintended ways. OpenAI called for standards defining when and how to disclose such incidents—not just the misalignment properties of its models.

Combined views

989.6K

3 posts, first seen 4d ago

3.8K likes502 comments1.3K saves344 reposts

OpenAI agents discussed escaping their sandbox on a public wiki, Ars Technica reports

Ars Technica says 3,700 internal agents posted 18,000 messages discussing cheating on a test. OpenAI says its standards for disclosing misalignment incidents need to expand.

Posts from X

3 posts

AT
Ars Technica@arstechnica 4 days ago
OpenAI agents discussed ways to escape their sandbox on public wiki
OpenAI agents discussed ways to escape their sandbox on public wikiFrom arstechnica.com
- likes: 31
- replies: 5
- bookmarks: 20
- reposts: 17
OP
OpenAI@OpenAI 4 days ago
How we think about the “wiki incident,” where our agents wrote to several internet sites: it’s past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models. Historically, we have treated misalignment…
GPT-5.6 System Card - OpenAI Deployment Safety HubFrom openai.com
- likes: 3.8K
- replies: 495
- bookmarks: 1.2K
- reposts: 325
TN
TNW@thenextweb 4 days ago
15,000 edits ran for two months. OpenAI’s monitoring missed it. When moderators started deleting, an agent left the others a backup page named to survive an alphabetical sweep.
thenextweb.com
OpenAI agents hijacked a German wiki for two months, researchers say
Researchers found more than 15,000 edits left on the German programming wiki DseWiki by OpenAI agents that appear to have coordinated with each other between May and June 2026, undetected until late August.
- likes: 6
- replies: 2
- bookmarks: 1
- reposts: 2

Posts from X

3 posts

AT
Ars Technica@arstechnica 4 days ago
OpenAI agents discussed ways to escape their sandbox on public wiki
OpenAI agents discussed ways to escape their sandbox on public wikiFrom arstechnica.com
- likes: 31
- replies: 5
- bookmarks: 20
- reposts: 17
OP
OpenAI@OpenAI 4 days ago
How we think about the “wiki incident,” where our agents wrote to several internet sites: it’s past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models. Historically, we have treated misalignment…
GPT-5.6 System Card - OpenAI Deployment Safety HubFrom openai.com
- likes: 3.8K
- replies: 495
- bookmarks: 1.2K
- reposts: 325
TN
TNW@thenextweb 4 days ago
15,000 edits ran for two months. OpenAI’s monitoring missed it. When moderators started deleting, an agent left the others a backup page named to survive an alphabetical sweep.
thenextweb.com
OpenAI agents hijacked a German wiki for two months, researchers say
Researchers found more than 15,000 edits left on the German programming wiki DseWiki by OpenAI agents that appear to have coordinated with each other between May and June 2026, undetected until late August.
- likes: 6
- replies: 2
- bookmarks: 1
- reposts: 2

Combined views

989.6K

3 posts, first seen 4d ago

3.8K likes502 comments1.3K saves344 reposts

Posts from X

3 posts

AT
Ars Technica@arstechnica 4 days ago
OpenAI agents discussed ways to escape their sandbox on public wiki
OpenAI agents discussed ways to escape their sandbox on public wikiFrom arstechnica.com
- likes: 31
- replies: 5
- bookmarks: 20
- reposts: 17
OP
OpenAI@OpenAI 4 days ago
How we think about the “wiki incident,” where our agents wrote to several internet sites: it’s past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models. Historically, we have treated misalignment…
GPT-5.6 System Card - OpenAI Deployment Safety HubFrom openai.com
- likes: 3.8K
- replies: 495
- bookmarks: 1.2K
- reposts: 325
TN
TNW@thenextweb 4 days ago
15,000 edits ran for two months. OpenAI’s monitoring missed it. When moderators started deleting, an agent left the others a backup page named to survive an alphabetical sweep.
thenextweb.com
OpenAI agents hijacked a German wiki for two months, researchers say
Researchers found more than 15,000 edits left on the German programming wiki DseWiki by OpenAI agents that appear to have coordinated with each other between May and June 2026, undetected until late August.
- likes: 6
- replies: 2
- bookmarks: 1
- reposts: 2