AI/ML applications have unique security threats. Project GuardRail is a set of security and privacy requirements that AI/ML applications should meet during their design phase that serve as guardrails against these threats. These requirements help scope the threats such applications must be protected against.

Saidot is a platform that helps you create and manage AI applications that are safe, ethical and transparent. With Saidot, you can: 1) Adopt AI policies from a library of templates, 2) Identify, mitigate and monitor AI risks, 3) Share AI reports with your stakeholders and 4) Learn how to use generative AI responsibly.

Comprehensive AI governance and compliance platform focused on empowering organizations with the risk management capabilities they need to adopt AI swiftly and responsibly.

Fast AI with Assurance, Integrity, and Reliability for comprehensive AI governance and oversight.

This paper addresses the complex and contentious issues surrounding AI governance, focusing on organizing overarching and systematic approaches for the transparency of AI algorithms, and proposes a practical toolkit for implementation. Designed to be highly user-friendly for both businesses and government agencies engaging in self- and co-regulation, the toolkit presents disclosure items in a list format, allowing for discretion and flexibility when considering AI algorithm providers, users, and risks.

FRAIA assesses the risks to human rights posed by algorithms and promotes measures to address them. It fosters dialogue among professionals involved in algorithm development or deployment. The client is accountable for implementing FRAIA to prevent uncertain outcomes of algorithm use. FRAIA mitigates risks of carelessness, ineffectiveness, or violation of citizens' rights.

QLECTOR LEAP is a cloud-based application that helps guide production processes in the same way the GPS supports your road trip. It uses historical data and AI to forecast production and suggest optimal measures when unplanned events occur. It combines machine learning methods specific for manufacturing and uses own AI technology to build and maintain data-driven digital twin of a factory.

Enterprise platform for governance, risk and compliance in AI. Regulatory compliance with global laws intersecting with AI.

GRACE is an AI governance platform that offers a central registry for all AI models, tools, and workflows for risk mitigation, compliance, and collaboration across AI teams and projects. It provides real-time compliance, transparency, and trust while promoting AI innovation. The platform works across all major cloud vendors and offers out-of-the-box frameworks for complying with EU AI Act, AI standards, data regulations, and AI frameworks.

An accessible framework (single A3 sheet) for governing the ethical development and use of AI - including the what (ethic), how (realisation principals), and submission of evidence for approval. Trustworthiness and bias are also covered

Zupervise is a unified risk transparency platform for AI governance.

An AI governance and risk management platform that helps organisations understand and manage the risks that come with AI, whilst navigating the emerging regulatory landscape

Model cards are the tool for transparent AI documentation. Model cards are essential for discoverability, reproducibility, and sharing. You can find a model card as the README.md file in any model repo. Under the hood, model cards are simple Markdown files with additional metadata.

A documentation-first, human-centered data collection project as part of the BigScience initiative. A geographically diverse set of target language groups were identified (Arabic, Basque, Chinese, Catalan, English, French, Indic languages, Indonesian, Niger-Congo languages, Portuguese, Spanish, and Vietnamese, as well as programming languages) for which to collect metadata on potential data sources. To structure this effort, an online catalogue was developed as a supporting tool for gathering metadata through organized public hackathons.