The AI systems we use are built on data. Data that are derived from the world we live in.
As Business Analysts (BAs) working as trusted advisors towards a better world, our job is still the same in its core. We still work with people trying to understand needs and the best ways to satisfy them. However, with AI, a new dimension is added to this work: We must now also safeguard the way AI systems are trained and the way they produce outputs, so that these outputs are aligned with the better world we all strive for.
For example, a new banking screening process should be fair to all. This was true before the advent of AI and it is still crucial in the AI era we live in. In both cases the advisor to the customer should pinpoint potential areas that could generate bias and eliminate them. The BA should work to clarify and establish the criteria used and how their application will result in better decisions. If in the past this was a set of COBOL programs in zOS or if in the future that will be done just through discussions with an AI agent, it matters not.
The true work is not about the technical details, but about the essence of what we do: Applying judgement where judgment should be applied. Putting safeguards where there are none. Understanding the big picture where others are focused on what is going on. Listening to stakeholders that do not have the opportunity to speak. Pinpointing potential issues in business processes that typically seem spot on.
Applying ethics in AI systems is a continuous process that needs to check the training data, the training process and the validation of the results as well. Doing so is difficult in many edge cases where the ethics of the process are not always clear. In any case we should also remember that ethics as a field is a field of strong philosophical debates for thousands of years. We will not solve everything overnight. However, we must try to do so overnight. Because AI is already here.
Difficult decisions are expected.
And it is in those decisions that advisors are needed.
Advisors trained in all those things we have forgotten for so much time.
Critical thinking. Empathy. Fairness. Saying “No”. Being unpleasant when everyone is having a good time.
Applying ethics requires clear mind and courage to support unpleasant things. To act or, when needed, not to. Not many people have the ability to manage ethical dilemmas and even less people have a solid philosophy to support the decisions they will make. Before we give AI systems more freedom to delve into critical areas where ethics play a major role, we should first clarify ourselves some things. Discussions we have been avoiding for centuries, must now be resolved within weeks.
If not anything else, the next years are expected to be extremely interesting.
And, hopefully, more ethical as well…
Related Projects
BA.AI 
Leave a comment