Ben Venzke’s Post

View profile for Ben Venzke, graphic

CEO at IntelCenter

I've been deep into operationalizing GenAI capabilities on IntelCenter's data now for five months. Three observations: 1) Periodically one of our GenAIs will write something and I'll look at it and think that can't be right. It must have it completely wrong. I'll then spend a good amount of time researching it only to discover that I'm the one who is wrong. (This just happened five minutes ago and thus this post.) 2) It's not always me. 😀 Sometimes I do find the GenAI is wrong. However, the number of times the GenAI is wrong is nothing compared to a green analyst or even a much more seasoned one who is incredibly overtasked. Also, a good percentage of those times the GenAI is wrong has more to do with how the data it's given is presented rather than its thinking being flawed. 3) There are many cases where it's not about being right or wrong. The GenAI can present thought provoking points that lead human analysts to explore new areas or inspire different ways of thinking. Here the GenAI response, whether spot-on, flawed or completely missing the mark is just acting as that initial spark and its contribution is valuable if it results in a valuable human output. Factor in that the GenAI is doing tasks in seconds that would take a team weeks to complete and you start to get the game-changing power this brings to the table. Everything said here though is predicated on the GenAI having access to quality data that is properly presented. Give the smartest GenAI in the world bad data and you will get bad responses. #genai #intelligenceanalysis #threatanalysis #osint #cai

To view or add a comment, sign in

Explore topics