Artificial Ignorance is a term that relates to machine learning and anomaly analysis, but as a concept, it can remind us to put in a process that ignores the noise of data overload and to focus on solving a problem, winning an opportunity, or opptimizing a business. In this post, we tie this to an important risk management concept: Identify, Measure, Monitor, and Report.
Parody of Artificial Intelligence?
Initially, Artificial Ignorance sounded like a parody that expressed a blissful retreat from the hype and fears of Artificial Intelligence. Synonyms for ignorance includes bewilderment and unlearned, portraying Artificial Ignorance is an automated way to do nothing.
It is not. In this case, ignorance is the act of ignoring unimportant data, allowing you to focus on what is important and necessary in order to create value.
Mitigating Risk Against the Unknown
Artificial Ignorance is a type of machine learning. It is mostly associated with monitoring log files and network traffic to spot security intrusions, identify root cause of system crashes, and other issues.
See the animated diagram below for this quick explanation:
- In the normal course of business, your organization generates data activity.
- A machine drops in and monitors the activity over time in order to learn what is normal.
- If something new enters and affects your environment, the machine identifies what is not normal, or anomalies, at very fine detail.
- It brings to your attention potential problems or perhaps, opportunities.
Since it is machine learning, issues are picked up that have not already been documented, mitigating risk against the unknown.
Trees for the Forest - Ignore and Optimize
The expression, “Can’t see the forest for the trees” is used to describe someone who is overcome by detail and fails to see the big picture. In this post, we consider the opposite, where value is derived by seeing specific trees for the forest. Depending on the business challenge, you may be looking for opportunites (trees with nuts) or problems to be fixed (diseased trees).
Mobile computing, real time analytics, and financial transactions from the edges of your world generates huge amounts of data in an instant. Depending on the situation, it may be optimal to ignore most of it (the forest) and focus on key events (some of the trees), optimizing the decision process.
Outside of the network infrastructure scenarios described above, the concept of Artificial Ignorance can be applied closer to the decision making process if you can get creative with the right tools and data assets in order to identify what is not normal (or what is special).
Deviation in Behavior and Compliance Based Events
A deviation in consumer behavior could flag a new business opportunity. Deviation of compliance based events could contribute key facts to a fraud examination and provide a compelling example to auditors of compliance monitoring and controls. Here are two examples:
- A trader action that deviates from what is considered normal trade activity could trigger a red flag. It may be worth monitoring that trader by additional and more descriptive means (chat, voice, review trade history, etc.).
- For trading strategy, the “something new” described in the animated diagram above, could be non traditional data like weather or location intelligence that puts a particular risk measure out of normal limits.
Summary: Identify, Measure, Monitor, and Report
A standard process in Risk Management is to identify, measure, monitor and report. Measuring, monitoring, and reporting comes to naught if relevant issues can’t be identified. There is often the desire to bring in more and more data but the tools need to be put that identifies what is important.
There is a lot of excitement and hype that comes with big data, real time analytics, artificial intelligence, the fintech revolution, and so forth. However, there is still a fundamental need to focus on business strategy (conceptualize and act on what is important) and data strategy (capture and identify what is important).
Depending on your business drivers, you probably don’t need all of the data that is being created. This is a good concept to keep in mind. As it relates to Artificial Intelligence, there could be very good reason to apply Artificial Ignorance first.