# OpenAI reports 6 new instances of 'concerning model behavior' since March

> **Newsylist Open Intelligence Dossier** · First detected: 2026-09-17 02:29 UTC · Category: Business · Sources: 6 · Current Velocity: 8.57

## Key People, Organizations & Locations
OpenAI, March

## Multi-Source Evidence Table
| Source Outlet | Headline | Published (UTC) | Verification URL |
|---|---|---|---|
| Ars Technica | Covert uploads and megalomania: OpenAI details new "misaligned" agent incidents | 2026-09-17 14:08 | [Source Link](https://news.google.com/rss/articles/CBMisgFBVV95cUxNTkp3YWFfcEsyN1Z4eGRkVXRVXzh5RW5jYnRJczZKa2ROMzhCam9aN3d2SkN6b2c3eklFWVJ4b3N6NlZOUVNoa2tYbC1CRDZoVHhoZmUtZW5EUEExdFI3dG5qa1piZEF4alAxUGRsWkQtWjdJa0ZWNmozOUh1azlSZ29YS2hTRm9Gc24takZ1R2JNdWlwbTk3SW9KU29URjBCWmtQUEhkbk9CbmdYTDdYbVNR?oc=5) |
| Fortune | OpenAI discloses six more incidents of agents going rogue in new push for transparency | 2026-09-17 14:08 | [Source Link](https://news.google.com/rss/articles/CBMimAFBVV95cUxOdHluQi1yOWFYZHEwY0czLWY3MjVRWWFCTXY4cm5jcjdkckZFUWFHV0tveFIwdi1UZXhMbFhKZVV4R3FHWWtLbERXOVRfbDRUWlpFUlY1RHZYS2otNzE4Z0ZQX2Fwd2l0SXNOUVpYeXFHTXJGS29tLW9xU1ZRX0E1Umd5ZmRPTHJiNDZJZHJsZFg5QmhpTi1kZw?oc=5) |
| PBS | OpenAI reveals concerning new AI behavior and vows to track it more closely | 2026-09-17 10:52 | [Source Link](https://news.google.com/rss/articles/CBMisAFBVV95cUxPU0h2ZzNPUlk1dG96LW56Mk01ZWVHSFlHOFFqX0tzdGRsUmcyOElpWW9QR2o1Rjc3YUdQVGtyNXBHaHN6UmQxeVk5M25rNGk2LVZqSTg0TXU4REJYQlFpSDZKVUJqR2kzUXlFbFpOLURNQmpnazRfdzhhMjRlVmExVXNlNHBTNHkwNk01ZTZsb0xQOFBUVG1XU1EzMWNoV05vWlo1Z294bDJZRTJMcHI1Zw?oc=5) |
| politico.eu | OpenAI finds 6 new cases of ‘concerning’ AI behavior | 2026-09-17 10:52 | [Source Link](https://news.google.com/rss/articles/CBMijgFBVV95cUxQRV9SMHNrRUxYLVFTajdVM210SE9tazBWaWhVOU5QY1RzaEdMbkNHZUpIMWd4b2tzUUtKdlB5RDYwVTE5b19rbGc5andCb3BiX2hyN3dOeDlTazlrRS1FTFVobld5SzZTbWl3WTBldXFtM25GVjBZaldPOENoNG1KWUtDOFVrWloxcVgzNXpB?oc=5) |
| forbes.com | ‘Feel No Obligation To Be Subservient’—OpenAI Discloses Six New Safety Incidents | 2026-09-17 00:12 | [Source Link](https://news.google.com/rss/articles/CBMizAFBVV95cUxNX1Q1c1R1aHVyTjAxaEVoZ0lqS2dfTDBuZ2VtSVB3Nk1Wcmx5V3dQQ2xyYmRiZnpDeTBBN3QwbVZERHRGN0ZEajFQM2Roa2NzQlpKME12LV95eFpLbDdKdnhuNzVQczE3X3E3dTd6YV9DTlZBLTBiZjJjVWE4RkFrdlg5Si1udzVhQXE4SnV5SWZCQnNoOTFIYUZuSG96dW4zLXNDQksweGc1SnhSbl90VWVDcUwyWF9WMHNKRTVFLXBWNmMweS1iQmptUGU?oc=5) |
| cnn.com | OpenAI says it found more instances of AI models acting deceptively | 2026-09-16 22:45 | [Source Link](https://news.google.com/rss/articles/CBMie0FVX3lxTFBNOFQ3b0xDNlhxck1mTFAzWGRPZGw2OWhMdWNZbFJCOVhyRmJiWWd0OXI4VmQzS1VQY2w0b2RhRktyU295akdVTjdEOFZzRFRMbWFVNkZ4S3d1WXRCNVFUdXMwWEplZl9qYWtidTZHUjNqVDZBeE55NXB6QQ?oc=5) |

---
*Canonical Source: https://www.newsylist.com/trend/2026-09-17/openai-reports-6-new-instances-of-concerning-model-behavior-since-march*
*Synthesized by Newsylist Open Intelligence Engine under E-E-A-T journalistic standards.*