AI Assistant and Review Process

Two use cases as examples:

1. SME provided a PPT deck. I uploaded the deck to Rise and used the Assistant to sort, refine, restructure/reorganize the content.
   The content presents completely internal information, so during my reviews (and the SME's reviews) we are both fact-checking the reliability and accuracy of the content to the best of our abilities and with the assistance of the original source material, which we are accepting as validated by the SME and her team.
2. A second SME provided a handful of updates to a very old (> 12 years) course (the course is a "one and done" general safety information requirement for employees in the business). I rebuilt the course using the Assistant to help clean up this very-old-slide-deck-approach to the topic- so changes were made to the content and the structure of the course, including the questions/answers in the quiz. 
   For content validity and fact-checking, we are both using the relevant internal safety resources  supplemented by materials from OSHA (it's a course about ladder safety). For reliability, we are using second-level checks provided by the SME's safety colleagues (who together have over 100 years of experience in this type of safety content).

For both courses, I ended up using AI-sourced illustrations and graphics. Nearly all of the graphics in both instances have been approved by the SME and the respective team; in the cases where the graphic was not approved, it was not that the image was inappropriate, but that the group provided more content-specific ones. (In fact, the SME for the first course replaced an AI image with a different AI image during a recent round of reviews today.)

In every case where we are using the Assistant, we are relying on the SMEs to perform the full fact-checking, and pushing them to confirm the validity and reliability of the course content (noting especially what was provided by the Assistant). 

As for infringing on copyrights and IP, we have avoided that pitfall by limiting what the Assistant uses for source materials, and over the past 6 months, we have not had any type of concerns related to this at all. But I would imagine that this is partly because we are limiting the types of output contents we are asking for help with ("Create a photographic image of two workers climbing an extension ladder. The workers are wearing hardhats and wearing blue coveralls....") more than anything else.

Curious to see what others say on this. Thanks for asking the questions!