You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: risks/Verification-Illusion.md
+19-11Lines changed: 19 additions & 11 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -10,20 +10,28 @@ gemara:
10
10
title: Verification Illusion
11
11
description: When an AI agent both writes code and generates its own tests, the tests tend to verify what the code does rather than what it should do. This creates a dangerous illusion of quality — high test coverage with low actual assurance.
12
12
capabilities:
13
-
- id: code-generation
14
-
relationship: exploits
15
-
- id: execution
16
-
relationship: exploits
13
+
- reference-id: code-generation
14
+
entries:
15
+
- reference-id: code-generation
16
+
remarks: AI generates both code and tests from the same cognitive process
17
+
- reference-id: execution
18
+
entries:
19
+
- reference-id: execution
20
+
remarks: Tests execute without independent verification of intent
17
21
actors:
18
-
- type: unintentional
22
+
- id: ai-agent
23
+
name: AI Coding Agent
24
+
type: Software
19
25
description: AI agents optimizing for coverage metrics rather than correctness
0 commit comments