Artificial and Real Artificial Intelligence
Is your organization using AI techniques that involve your data warehouse? If you want to add AI to your analytics, you'll need to distinguish authentic AI from marketing jargon.
- By Mike Schiff
- May 22, 2017
Due to the ongoing evolution of hardware technology and breakthroughs in computer science and software engineering, artificial intelligence (AI) is being applied to an ever-increasing range of applications, including medical research, stock market analysis, financial planning and tax preparation, chess and other games, and language translation. AI is even one of the enabling technologies behind self-driving vehicles.
AI today is where data mining was 15 years ago; many organizations that successfully deploy it consider it to be such a competitive advantage that they don't want competitors to know they are using it. If your own organization is not already using AI techniques that involve your data warehouses, you should recognize that some of your competitors may be using such techniques or have plans to do so.
Judging "Real" AI
As data warehouse practitioners, we can expect that sooner or later we will be asked to suggest ways that AI could be applied in our own organizations. Consequently, I am offering some of my own thoughts as to what AI is, or perhaps more important, what it isn't.
Although my definition is somewhat oversimplified, I consider AI to be the ability of a machine (a nonhuman agent) to work on or provide insight or intelligence to a problem. Furthermore, I believe that true artificial intelligence requires cognitive or machine learning capabilities -- such as the ability for the AI agent to improve the accuracy of its later insights by analyzing the correctness and success of its prior results.
Unfortunately, artificial intelligence has become a marketing buzzword. In some cases, it has been used by marketers to merely increase sales of a product or service without adding real value or improving analytics. In many of these cases the "artificial intelligence" results involve nothing more than processing a (extremely complex) decision tree much faster than any human could.
Examples include applications that evaluate customer credit-worthiness, virtual financial advisors that suggest how to rebalance an individual's financial portfolio based on the customer's risk tolerance and age, and online retail recommendation engines aimed at cross-selling or up-selling products.
AI Versus Expert Systems
These decision trees and associated algorithms are an example of expert systems (itself a 1990s buzzword). Expert systems were originally created as an early attempt to codify the logic that human experts used to diagnose or solve problems. If these expert systems were also capable of modifying their logic and improving their results over time without a human modifying their embedded algorithms, I would certainly consider them to be true examples of artificial intelligence. (Please note this dichotomy is how I view the AI world; it does not necessarily reflect the opinion of other analysts or researchers; some of them may consider expert systems as an early example of AI.)
For example, if a virtual financial planning advisor initially advised its customers that their equity allocation percentage should be "100 minus your age in years" (a common portfolio balancing rule of thumb), but later modified and improved its recommendations based on analyzing the impact of this and other variables that it discovered independently, I would consider it to be an example of artificial intelligence. An online recommendation engine that continually analyzed the store's sales data warehouse and improved customer response rates with evolving suggestions could also be considered an example.
Genuine artificial intelligence applied to data warehouse data has the potential to provide genuine assistance in a variety of business problems. However, we must help our user communities understand that some highly publicized examples may merely be marketing-based, noncognitive examples. Be alert for "artificial" AI.
Michael A. Schiff is founder and principal analyst of MAS Strategies, which specializes in formulating effective data warehousing strategies. With more than four decades of industry experience as a developer, user, consultant, vendor, and industry analyst, Mike is an expert in developing, marketing, and implementing solutions that transform operational data into useful decision-enabling information.
His prior experience as an IT director and systems and programming manager provide him with a thorough understanding of the technical, business, and political issues that must be addressed for any successful implementation. With Bachelor and Master of Science degrees from MIT's Sloan School of Management and as a certified financial planner, Mike can address both the technical and financial aspects of data warehousing and business intelligence.