Skip to main content
Skip table of contents

How to use Saidot Library

Policy Library

Saidot Policy Library helps organisations and individuals to understand the variety of different AI, data and digital policies contained in the Library. The Policy Library contains an ever-growing collection of policies, policies ranging from laws to standards, frameworks, guidelines, and best practices.

Step 1. Find policies

The main page of the Policy Library allows you to browse and view different policies. With the help of filters you can find policies from specific regions, sectors, and industries. You can furthermore filter policies based on their type as well as owner.

image-20241119-121836.png

Step 2. Read policy overviews

Selecting and clicking relevant policy allows you to familiarise yourself with the policy. The policy overview contains useful overview information on the policy, it’s scope, as well as it’s key requirements - in some cases you can also find information about relevant penalties under the policy. Through the ‘reference’ button in policy overview you can also easily navigate to the official source of the policy.

image-20241119-122033.png

Step 3. Familiarise policy roles

In most cases, you can find information about the roles regulated under the given policy - this information is handily provided in the ‘Role’ tab, next to the overview.

image-20241119-122115.png

Step 4. Understand policy coverage

In some cases, in addition to the regulated roles, you can find information about the different categories of policy coverage regulated in the policy. ‘Coverage’ tab records the different categories of policy coverage - whether systems or activities - that are regulated under the policy.

image-20241119-122215.png

Step 5. Dive deep into policy requirements

For selected policies, you may find policy deep dives. ‘Requirements' tab provides you information of the requirements in the policy, in a structured manner with understandable descriptions. in ‘Requirements’ you can navigate to the different requirement sections and familiarise yourself with the different requirements of the policy. Role indicates to you to whom the requirement applies, and references informs you from where in the policy the requirement can be found.

image-20241119-122256.png

Risk Library

The Saidot Risk Library helps organisations understand and manage over 100 AI-related risks. It provides detailed descriptions of each risk, how they manifest, and their impact, along with practical mitigation strategies. Users can filter, sort, and follow specific risks for updates.

Step 1. Find risks

Find relevant risks by using the filters. Risks can be filtered by:

  • AI System: Filter risks based on the specific AI system you're working with.

  • Type of Risk: Select from categories such as legal, cybersecurity, environmental, technical, trust, fundamental rights, privacy and data protection, societal, third-party, business, and health and safety.

  • Industry: Narrow down risks relevant to your specific industry, such as media and marketing as well as banking and finance.

Specific risks can be followed to stay updated and informed about relevant developments. Risk view can be sorted by most recent, alphabetical order, or followed risks.

Risk Library Macbook.png

Step 2. Understand risks and mitigations

Each risk entry provides a comprehensive description, including:

  • definition of the risk

  • explanation of how the risk can occur

  • potential consequences and impact of the risk

image-20241119-122546.png

Step 3. Analyse potential mitigations

Each identified risk comes with mitigation strategies, offering clear and practical guidance on effectively addressing and managing specific risks. This helps organisations implement robust risk mitigation plans and reduce the impact of potential threats.

image-20241119-122623.png

Model Library

Saidot Model Library helps organisations and individuals navigate and analyse AI models with confidence. The library provides essential insights through simplified model analysis, capability assessments, and provider-identified risks, empowering teams to make well-informed AI deployment decisions. Each model entry includes a clear analysis of terms of service, usage restrictions, and content policies, ensuring teams and users understand both the opportunities and limitations before implementation. Through structured documentation of technical capabilities and usage constraints, the library serves as a trusted resource for compliance management throughout the AI governance lifecycle.

Step 1. Find models

The main page of the Model Library allows you to browse and view different models. With the help of filters you can find policies from specific vendors, types, industries, tasks they perform, architecture, and region.

image-20241119-124933.png

Step 2. Explore the models

Click on any model to explore the model analysis. The model page opens up to an overview of the technical details, model evaluations, safety resources and research references that provide more details about the model.

image-20241119-125046.png

Step 3. Analyse technical evaluations

Models come with capability and safety evaluations offering insights for responsible AI deployment, helping Saidot’s users manage risks and optimise resource allocation.

image-20241119-125116.png

Step 4. Understand model-related safety features and risks

Each model entry provides a comprehensive safety analysis, including:

  • Risks

  • Restrictions

  • Provider safety features

image-20241119-125149.png

Step 5. Explore the product page directly from the library

The Model Library streamlines access by connecting you directly to provider resources and documentation in one place. Instead of scattered research, you can quickly access official vendor information and materials through a single, organised hub.

image-20241119-125224.png

Step 6. Keep track of models and related risks

The Model Library connects models with their related risks, allowing you to track changes by following both models and their risk pages. This helps you stay informed about important updates and emerging concerns that could affect your AI systems.

image-20241119-125317.png

JavaScript errors detected

Please note, these errors can depend on your browser setup.

If this problem persists, please contact our support.