ACE Platform - EU AI Act Compliance Statement

1. Executive Summary

This document describes how the ACE (Agentic Context Engineering) platform complies with the European Union Artificial Intelligence Act (EU AI Act, Regulation 2024/1689).

2. System Description

2.1 What is ACE?

ACE is an AI-powered pattern learning platform for software developers that:

1. Receives execution traces (coding task descriptions, steps taken, results)
2. Analyzes traces using AI (Anthropic Claude) to identify patterns
3. Stores learned patterns in a structured "playbook"
4. Retrieves relevant patterns for future coding tasks

2.2 AI Components

2.3 Value Chain Position

┌─────────────────────────────────────────────────────────────────┐
│                    EU AI ACT VALUE CHAIN                        │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  GPAI MODEL PROVIDER     AI SYSTEM PROVIDER      DEPLOYER      │
│  (Anthropic)             (Code Engine/ACE)       (Users)       │
│       │                        │                    │          │
│       │  Claude models         │  ACE platform      │          │
│       │  (Sonnet, Haiku)       │  (API + SDK)       │          │
│       │                        │                    │          │
│       ▼                        ▼                    ▼          │
│  ┌─────────┐            ┌───────────┐        ┌─────────┐       │
│  │ GPAI    │ ─────────▶ │ ACE       │ ──────▶│ End     │       │
│  │ Models  │ integrated │ Platform  │  used  │ Users   │       │
│  └─────────┘   into     └───────────┘   by   └─────────┘       │
│                                                                 │
│  Anthropic's       Code Engine's        Customer's             │
│  obligations       obligations          obligations            │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

3. Risk Classification

3.1 Classification Analysis

Under Article 6 of the EU AI Act, AI systems are classified based on their intended purpose and potential impact.

3.2 Why ACE is NOT High-Risk

ACE does not fall under Annex III high-risk categories because:

3.3 Safeguards Against High-Risk Use

To ensure ACE is not used for high-risk purposes, we implement:

1. Acceptable Use Policy - Explicitly prohibits Annex III uses
2. Terms of Service - Legally binding prohibition on high-risk use
3. Technical Controls - No features designed for HR/employment use
4. Monitoring - Usage patterns reviewed for policy compliance

4. Transparency Compliance (Article 50)

4.1 AI Disclosure

Users are informed that they are interacting with an AI system through:

4.2 AI Processing Information

Users are informed about:

• Which AI models process their data (Claude Sonnet, Claude Haiku)
• What decisions AI makes (pattern retention, similarity matching)
• How to override AI decisions (management interface)
• How to disable AI processing (configuration settings)

5. Human Oversight (Article 14)

5.1 Oversight Mechanisms

Users maintain oversight through:

5.2 Effective Oversight Design

The system is designed so that:

• No AI decision is irreversible
• Users can inspect all AI reasoning (patterns include evidence)
• AI suggestions can be ignored or overridden
• Human judgment is final in all cases

6. Data Governance (Article 10)

6.1 Training Data

ACE does NOT train AI models. We use pre-trained models from Anthropic (Claude). Anthropic is responsible for their training data governance.

6.2 User Data Processing

For user-submitted data, we ensure:

7. Technical Documentation (Article 11)

We maintain documentation covering:

• System architecture and data flows
• AI component specifications
• Security measures and controls
• API specifications
• Operational procedures

8. Record-Keeping (Article 12)

We maintain comprehensive logs:

9. Accuracy and Robustness (Article 15)

10. Cybersecurity (Article 15)

11. Provider Obligations Summary

As an AI System Provider under the EU AI Act, Code Engine:

12. Deployer Guidance

If you integrate ACE into your own products, you may be a "deployer" under the EU AI Act with your own obligations:

12.1 Your Potential Obligations

12.2 Resources

• EU AI Act Full Text
• EU AI Act High-Level Summary
• Article 26: Deployer Obligations

13. Contact for Compliance Inquiries

For questions about EU AI Act compliance:

• Email: compliance@code-engine.app
• Subject: "EU AI Act Inquiry - ACE"

For compliance documentation requests (enterprise customers): legal@code-engine.app