Gauging interest in leveraging AI to improve Risk Solutions and Custom Responses
We have heard from some of our FedRAMP authorized clients about the desire to retain some of the extra detail captured in your audited SSP.
We are developing AI-powered solutions to help adopt Risk Solutions while also retaining the detail captured in your SSP.
What is your interest in this functionality?
What, if any, are your reservations about AI usage?
Rest assured, we recognize the sensitivity of your FedRAMP data. Our solutions will fully align with FedRAMP’s stringent security and privacy requirements, ensuring that your data is handled in a way that meets the applicable controls.
-
My key question and concern is what the AI would use as a model? How is the information within the LLM stored or integrated based on the SSP language and is there a way to exclude information or details from being absorbed into the LLM? I would want to see a proposed architecture and data protection model before we could even consider it. What information would be used to train the model?
The challenge with utilizing an AI in this is that “smoothed” or rewritten language over time and across work products become “samey” or generic, and the author's and organization's voice or tone could be lost and altered into an amalgam. Many AI-crafted documents also suffer from odd hallucinations and repetition.1
Please sign in to leave a comment.
Comments
1 comment