In the past, quarterly governance meetings seemed more than adequate. Reports would be created, risk analyses conducted, and decisions based off of everything that we had known at that time.
Artificial intelligence doesn’t operate within the same temporal frameworks. Artificial intelligence works continuously.
The idea of real-time governance dashboards has emerged from this disconnect. Rather than only looking at historic reports, leaders can access “live” data for how their predictive models are performing, compliance metrics, and early alert systems for emerging risk.
For example, if a loan approval model is showing bias across certain demographic categories, would you rather have a monthly or a “present” notification of that occurrence? You would probably want to have real-time, so that you can take immediate action.
Check out: Finding Strategic Direction in an Era of Constant Noise
At the same time, it is important to understand that having more information does not always lead to better decisions. If we provide too much information, it can create a cluttered dashboard where all metrics compete for our attention and represent a different type of failure.
Therefore, the challenge will be identifying what is useful and what is not; in fact, determining which of the selected metrics are useful will also create another level of confusion.
Additionally, it will be important to establish a framework for interpreting anomalies and determining the context for a particular occurrence (e.g., is an anomaly an actual issue, or is it just a spike due to temporary fluctuations?).
There is no denying that a change is occurring. Governance now is moving from episodic to continuous. Governance boards no longer simply review the outcomes of governance processes; they observe real-time governance systems in motion.
Ultimately, this suggests that “what is governance” will be less about reviewing past actions and more about being aware of current actions.
