Artificial Intelligence: An Inconvenient Data Truth
We have created a massive data problem for ourselves and we think that by using artificial intelligence and creating more algorithms we will get it all sorted.
The data, the sensors, the rules and the optimization algorithms will seamlessly make the best decision to produce the right outcome at the right time.
There are multiple problems with this approach. Artificial intelligence is not:
1. Great at picking up on context clues and understanding the full picture.
2. Effective at knowing what to do in new situations. Artificial intelligence can guess based on data that it has about similar scenarios, but is it really any better at guessing than the human mind.
3. Looking at risk in the same way that a human would in the same circumstance.
4. Appreciative of the role of timing in any decision. Decisions don’t happen in a vacuum. Many decisions are being made independently at the same time and an artificial intelligence may not be programmed to recognize relevant factors in the same that humans are.
Relying on artificial intelligence to solve problems that don’t always have a logical or rational answer may not be the best strategy for finding the best long term solution.
Yet, we plod along trying to apply artificial intelligence to more situations and expect that it will solve problems better and save us time from dealing with all the data.
The inconvenient reality about data is that it is an ever expanding set. More data is collected everyday than is discarded and while storing the data is relatively inexpensive, using the data is potentially very costly.