Bring experience to the AI debate

Where was all this excitement about humans in, on or out of the loop when we installed advanced process control in the refinery? Or when we added deep learning modules on top?

That experience is a useful starting point for today's AI debate. It prompts a practical question: what authority are we giving the system, in which operating conditions, and who remains responsible for the result?

Labels about human oversight become useful once they describe actual work. A person approving every recommendation, an operator supervising a process and a manager reviewing exceptions have different jobs. Each needs information, time and the ability to intervene.

I would take one proposed AI application and follow a difficult case through it. What does the system do? What reaches a person? What can that person check? What happens when the conditions change?

Working through those questions with a real application builds the experience leaders need to judge the proposal. Familiarity with earlier automation gives us questions worth asking. Hands-on work with the current system gives us evidence for the answers.

Stay composed, get the experience and lead confidently.