Using ChatGPT to Evaluate Whole Logical Diagrams: Promising First Results
Valid causal analysis is the key to valid conclusions and well grounded decisions. AI does now provide us with the ability to vastly improve our causal analysis and help us make sounder decisions.
During the past few months I’ve experimented with using ChatGPT to evaluate and improve logic in complicated cause-effect diagrams. I discuss this in length in my recently published training course which is available here; how it can speed up the structuring and scrutiny of logical analysis and help improve the validity of our logic.
The latest addition to GPT 4.0 is the ability to let the model read diagrams. I have now tested this new functionality and the results are very promising. Below, I’m sharing a prompt and the model’s response, using a simple, but still a multi-level cause-effect diagram.
In the context of management, and in fact decision-making in general, this is great news. It means that using a structured approach, such as the Logical Thinking Process methodology, which until now has for many been too difficult and time-consuming to do, is now much more practical. And since valid causal analysis is the key to valid conclusions and well grounded decisions, AI does in this way provide us with the ability to vastly improve our causal analysis and help us make sounder decisions.
In this test I’m using the GPT 4.0, which is the paid version of ChatGPT, and a diagram created in Lucidchart, exported to a pdf file.
