This is a real certificate generated by AICVS from analysis of example AI code — illustrative SAMPLE, not legal certification.
AICVS Technical Evidence Certificate

EU AI Act code-evidence report

Subject: credit_scoring.py · Python · scanned 2026-07-08 · layers: Regex+Statistical+AST+Semantic · 23 lines
29
/ 100

Verdict

FAIL Classification: AI_GENERATED

Significant AI compliance signals detected in 'credit_scoring.py' (score 29/100). AICVS found 5 signal(s) collectively suggesting high AI involvement. Critical signals (0): none. High-confidence signals (2): AI-002 (OpenAI API integration), AI-002 (OpenAI API integration). EU articles implicated: Art.11, Art.13, Art.15, Art.9. IMPORTANT FOR REVIEWERS: This identifies specific, explainable signals — it is NOT a binary verdict. Review each finding, ask the author to explain their code, and use this as supporting evidence alongside oral examination.

EU AI Act articles implicated

Art.11Technical documentation (Annex IV)
Art.13Transparency & information to deployers
Art.15Accuracy, robustness & cybersecurity
Art.9Risk management system

Findings (5)

RuleSeverityFindingEU AI ActLine
AI-002HIGHOpenAI API integration (GPAI system)Art.9,Art.131
AI-002HIGHOpenAI API integration (GPAI system)Art.9,Art.1312
STAT-002MEDIUMLow identifier variance (σ²=3.9, 0% short names)Art.11
STAT-009LOWRegular indentation step pattern (delta entropy=0.93)Art.11
AST-004MEDIUMVery low cyclomatic complexity (0.5 decisions/function avg)Art.9,Art.15

Evidence chain (SHA-256 Merkle)

Six tamper-evident steps bind the file content, findings, and article mapping into a single sealed digest. Any change to the file or findings changes every hash below.

1
Content ingest
UTF-8 content SHA-256; file=credit_scoring.py
110a59facfc54b6afe1e150fa6dae43c7668262d9575a1f2d8909bf3e7c2251a
2
Identity binding
Scan ID bound: 414388a3…
703f1e781b02e45115a10b966cc392885cbd25ecd0f88eaf461b8de1fbd8aec4
3
Provenance & findings
5 signal(s) fingerprinted
586fc8ffa6a4334b912682e1ddb3a763abee1c2037fb8e5915be18c44fc38549
4
EU AI Act mapping
Articles: Art.11, Art.13, Art.15, Art.9
91cfe6c1174499d06585d05382d9cccdfb273ea56443e8177bddbfd808594f25
5
Merkle seal
AICVS-1.11.0
f9b7566b4253c299f846544290b4796706df2ef884b009679a80439caafb1b07
6
Certificate digest
Public verification hash
ac490a21fa02839f40b74e2aef96a281cae9bfc83eaba1462695570bfb3a699d
Merkle rootf9b7566b4253c299f846544290b4796706df2ef884b009679a80439caafb1b07
Certificate hashac490a21fa02839f40b74e2aef96a281cae9bfc83eaba1462695570bfb3a699d
File SHA-256abc638a08c16c40620b10c19a8fc02a145c13875679085021a77a966466bda3a

How verification works

On a saved scan, anyone can confirm a certificate without contacting AICVS: the public endpoint GET /api/v1/scans/{scan_id}/verify returns the certificate hash and status, and /verify/{scan_id} checks the SHA-256 chain and RFC 3161 timestamp. Because the chain is tamper-evident, a modified file or altered finding produces a different digest and fails verification.

This sample was generated from AICVS's real analysis of example credit-scoring code; the scores, findings, article mapping, and hashes above are genuine engine output, not mock-ups. It is illustrative and not saved for public verification. AICVS output is readiness support based on the code provided — not legal advice, not a conformity assessment under Article 43, and not legal certification. Review with qualified counsel.

Generate your own — free

Scan one of your own AI files in the browser (no signup), or create a free account to save a persistent, publicly verifiable certificate and scan whole projects.

Scan your AI code →