Builders and buyers of AI systems are required to test and show that their systems comply with legislation – on safety, discrimination, privacy, transparency, and accountability. This talk covers recent regulation in this space, limitations that current Generative AI models have, and an automated testing framework that mitigates them.We describe the open-source LangTest library, which can automate the generation and execution of more than 100 types of Responsible AI tests. We then introduce Pacific AI, which provides a no-code interface for this capability for domain experts, as well as automating many of the best practices on how these tools should be used.
Builders and buyers of AI systems are required to test and show that their systems comply with legislation – on safety, discrimination, privacy, transparency, and accountability. This talk covers recent...
Side-by-side viewing and comparing different annotations made by various annotators on text documents present several challenges. Firstly, aligning the annotations for accurate comparison can be difficult, especially if the annotators...
Overall, de-identification in today’s data-driven world is a critical practice that helps balance the benefits of AI and big data with the need for privacy and compliance, facilitating both technological...
In industries like healthcare, in which regulatory-grade accuracy is a requirement, human validation of model results is often a critical requirement. While models handle the legwork, the Generative AI Lab...
Demos
See More