May 26, 2024 7:12 PM
Replying to Keith Novak
...
There are a lot of ways AI can support risk management. The question is what problem you are trying to address and whether AI is the right solution for the application.
LLMs can help compare your project to case studies to identify common issues. I think AI has the best potential yet for making documented lessons learned actually of some real value.
Image recognition algorithms are already used in various applications to detect potential threats. Is there a risk on the project which could be detected visually if there were cameras and/or radars monitoring some important work area?
Machine learning algorithms can review very large datasets for patterns us mere humans would never find. AI can be used to find trends and then provide notifications of something significant like an incredibly powerful control chart.
There are also negative aspects like cost and trusting the data. Some problems are immensely complicated and involve massive data requiring supercomputers. Other problems can be mitigated with a caution sign. Constantly being watched is perceived differently in the context of a nuclear reactor core, vs. an integrated laptop camera in someone's home office.
I personally use AI embedded in a variety of tools such as grammar aides, search engine type tasks, enabling applications in MS Office, and I'm learning Minitab for statistics. None are mind-blowing technology in today's age, but they do enable capabilities that were much more labor intensive before assistance from AI. That can change the cost benefit ratio from not worth the effort, to easy and highly effective.