1. Microsoft recently unveiled new Responsible AI features aimed at enhancing trust in AI systems during the AI Tour in Mexico City. With a focus on privacy, safety, and security, these features include risk evaluations and content safety tools. The initiative emphasizes transparency and data control, ensuring responsible AI development.2.

New Responsible AI Features for Building Trustworthy AI
Microsoft recently announced exciting updates aimed at enhancing trust in AI systems. These features focus on privacy, safety, and security.
What’s New?
During the Microsoft AI Tour in Mexico City, the company introduced its “Trustworthy AI” initiative. This approach emphasizes a commitment to responsible AI development.
“Every AI innovation at Microsoft is grounded in a comprehensive set of AI principles, policies, and standards.”
This initiative includes foundational commitments like the Secure Future Initiative and AI Principles. These frameworks ensure users have control over their data, whether at rest or in transit.
Major Updates
Microsoft unveiled four new capabilities in public preview to enhance the evaluation of generative AI applications:
- Risk and safety evaluations for indirect prompt injection attacks.
- Evaluations for protected material (text).
- Math-based metrics like ROUGE, BLEU, METEOR, and GLEU.
- A synthetic data generator for non-adversarial tasks.
These tools provide a systematic approach to evaluate AI outputs, moving beyond intuition and sporadic feedback.
What’s Important to Know
A significant change involves migrating evaluators from the promptflow-evals package to the new Azure AI Evaluation SDK. Users should prepare a migration plan to avoid errors related to missing inputs.
“We’re transparent about where data is located and how it’s used, and we’re committed to making sure AI systems are developed responsibly.”
For developers in regulated sectors, data protection remains a priority. New features include:
- Confidential inferencing.
- Azure OpenAI Data Zones.
These features address the challenges of processing sensitive data in the cloud while ensuring encryption at all times.
Conclusion
Microsoft’s commitment to responsible AI is evident in these new features. By prioritizing privacy, safety, and security, they aim to build trust in AI systems. Developers and enterprises can now leverage these tools to create applications that meet high standards.
From the Microsoft Developer Community Blog