Autodesk Embeds AI Transparency Cards Directly into Assistant

Autodesk has integrated its AI Transparency Cards into Autodesk Assistant, allowing users to access detailed explanations of AI features without leaving their current workspace. This update responds to customer feedback that trust questions often arise mid-task, requiring quick, reliable answers about what a tool does and how it handles data.

The cards, first introduced in 2024, serve as a standardized reference for AI capabilities. By embedding them into the assistant, Autodesk aims to reduce friction for professionals who need to verify safety and transparency constraints before running models on client projects. The move shifts the focus from passive documentation to active, contextual support.

Transparency cards explain model origins

Modeled after nutrition labels, each card breaks down an AI feature’s function, the underlying models, and training methods in plain language. The updated version includes new sections on data handling, customer choice, and safeguards, addressing gaps identified in previous iterations. This structure helps users understand not just what the AI can do, but the boundaries and protections that govern it.

Research conducted with Deloitte indicated that while initial cards increased perceptions of transparency, they also raised user expectations. Customers wanted clearer information on how data is processed and what oversight mechanisms exist. The new design responds to this by explicitly detailing the sources of training data and the specific safeguards applied, such as data minimization and access controls.

Immediate access supports decision making

Users can now ask the assistant specific questions about data processing, limitations, or human oversight, and receive answers that cite the relevant transparency card. This allows for a deeper dive into the source material if the summary is insufficient. The goal is to settle concerns quickly, preventing potential delays in project workflows that might occur while searching for external documentation.

For many professionals, a clear, one-page explanation is enough to resolve doubts about using AI on sensitive client work. By placing this information directly in the tool, Autodesk removes the need to switch platforms or consult separate trust centers. This integration supports a more confident adoption of AI tools, as users can verify compliance and safety in real time.

Data usage remains feature dependent

A key trade-off in AI adoption is the uncertainty surrounding data usage. According to information reported by GN technics/ai, Autodesk clarifies that it does not share customer data with third parties for their model training. However, for some internal features, aggregated and de-identified customer content may be part of training datasets, limited to non-sensitive elements like industry standards.

This distinction is critical for users to understand, as the application of AI varies by feature. The transparency cards now explicitly state whether a specific tool uses customer data for training and what safeguards, such as anonymization or secure processing environments, are in place. This clarity allows users to make informed decisions based on their specific data privacy requirements and compliance needs.