EU AI Act-Compliant Open-Source Time-Series Forecasting Package
A new open-source Python package, spotforecast2-safe, implements Compliance-by-Design for time-series point forecasting in safety-critical environments. Unlike existing compliance tools that operate externally as scanners or runtime layers, this library embeds requirements from the EU AI Act (Regulation 2024/1689), IEC 61508, ISA/IEC 62443, and the Cyber Resilience Act directly into API contracts, persistence formats, and CI gates. The approach enforces four non-negotiable code rules: zero dead code, deterministic processing, fail-safe handling, and minimal dependencies, alongside process rules like model cards and executable docstrings. The package targets industries where forecasting errors could have severe consequences, such as autonomous driving, industrial automation, and healthcare. By integrating compliance at the library level, spotforecast2-safe aims to reduce certification overhead and ensure regulatory adherence from development through deployment.
Key facts
- spotforecast2-safe is an open-source Python package for time-series point forecasting.
- It uses a Compliance-by-Design approach for safety-critical environments.
- It embeds EU AI Act, IEC 61508, ISA/IEC 62443, and Cyber Resilience Act requirements.
- Existing compliance solutions operate outside the library as scanners or runtime layers.
- Four code rules: zero dead code, deterministic processing, fail-safe handling, minimal dependencies.
- Process rules include model cards and executable docstrings.
- Targets industries like autonomous driving, industrial automation, and healthcare.
- The package aims to reduce certification overhead and ensure regulatory adherence.
Entities
—