|
|
|
|
Tuesday, November 24, 2009
|
ISSUE 45
|
|
|
When will machines replace analysts?
Philip J. Annibale, Mercyhurst College
Its 8:00 on a crisp, cool Friday morning with the birds chirping and the sun rising. You walk into your office building with a cup of warm decaf in one hand and your leather briefcase in the other. As you place your briefcase next to your office desk, you discover a letter from your boss informing you that you and your duties are about to be replaced by a machine.
Stunned? Confused? Frustrated? Artificial intelligence (AI) is evolving and expanding at such a rapid pace that machines could replace the daily tasks performed by analysts by the year 2030.
You may be thinking to yourself, “Machines will never perform my job because of the many unique tasks I complete everyday.” This notion, however, is gradually becoming obsolete. AI now has the ability to simulate a variety of characteristics of the human brain. With readily available software, machines are starting to perform certain cognitive skills including problem solving, adaptive learning, and the ability to understand a variety of languages.
In addition, machine intelligence can function 24 hours a day, does not require a salary (or benefits), and can store and process greater amounts of data more quickly than humans. If your palms are sweaty and you’re already at the edge of your seat then sit back, take a deep breath, and look even closer at the evolution of AI.
Machines vs. analysts in core competencies
When compared to the various abilities and skills of human analysts, AI is beginning to simulate analytical tasks more effectively and efficiently. To illustrate, the following graph shows the advancements of AI in relation to the abilities and skills of human analysts (Moore, David, and Lisa Krizan. “Intelligence Analysis: Does NSA Have What it Takes.” Cryptologic Quarterly, vol. 20, nos. 1/2 (Spring/Summer 2001): 1-33. ), as well as future predictions of AI’s potential.
| Computer proficiency in analyst characteristics |
Research proof of concept
|
Full computer proficiency as good as analysts
|
Communicating
|
1990
|
2015
|
Information ordering
|
1994
|
2004
|
Pattern recognition
|
1995
|
2010
|
Teaming and collaboration
|
2020
|
2020
|
Reasoning
|
2030 |
2030 |
Info gathering & manipulation
|
1985
|
2002
|
Basic literacy
|
1990
|
2010
|
Research
|
1994
|
2010
|
Foreign language
|
1995
|
2015
|
Project/process management
|
1998
|
2020
|
Critical reasoning
|
2030
|
2030
|
| Expression |
2030
|
2030 |
Information ordering and foreign language proficiency are two functions once performed by analysts that are now partially executed by machines. The development of the Web provided an opportunity to speed up growth within the AI field. For example, search engines such as Google can now systematically arrange information based on keywords, a task once executed by analysts.
In addition, the emergence of the internet also contributed to the expansion of multilingual translation software. Since 1995, numerous companies began to offer free online trial sessions, along with online memberships ranging from $20 to $180, of software capable of translating many kinds of foreign text documents with a click from the mouse. Foreign language proficiency software is also expanding to not only translate, but also teach foreign languages.
Three analytical functions of analysts that are currently becoming obsolete because of the advancements of machine intelligence include: pattern recognition, research methods, and basic machine literacy.
Pattern recognition
Pattern recognition is the ability to identify similarities and irregularities by using algorithms (a set of computer procedures) to classify data. With this advancement, pattern recognition is becoming a valuable tool in the field of biometrics, or the analysis of identifying features in humans based on physiological or behavioral characteristics.
According to Dr. Peter McOwan, “Biometrics is currently a key area of research worldwide. Current biometric systems most commonly identify individuals through some unique characteristic for example by their iris, retina, voice pattern, fingerprints or facial features as well as the pressure patterns of hand-written signature.” (“Artificial intelligence to increase security of online shopping and banking,” Queen Mary, University of London. Department of Computer Sciences. http://www.qmw.ac.uk/poffice/nr270803.shtml, October 2003 )
Examples of technological devices that use AI include the Panasonic CCD and Sony cameras which are able to analyze facial dimensions, the Veridicom Fingerprint Sensor with the ability to “read” and classify individual fingerprints, and the Pulnix HandGrabber capable of matching hand configurations to individuals. Because of the September 11th terrorist attacks on the World Trade Towers, the U.S. Government is about to roll out new biometric technology that could enable the Department of Homeland Security to check the identity of every visitor to the U.S., an estimated 440 million annual visitors!
Research methodology
Implementing research methods is another analytical role currently affected by machine intelligence. Proprietary algorithms on the internet are now able to identify a pattern in a person’s wants and needs and then offer suggestions for products that correlate to the individual’s desires. Amazon.com and mysimon.com simulate virtual shopping malls that do the “shopping” for you. Simply complete an online questionnaire signifying your desires and specifications in a product and allow the search engine to research and report to you products that match your criteria.
A machine’s ability to perform and analyze research offers several advantages in producing more dynamic intelligence. For example, an analysis of the geopolitics in Albania as well as an examination of Albania’s economic conditions could trigger research that suggests the possibilities of the next ruling political party in Albania. The ability of machines to research and present related information according to an individual’s search interests is also starting to develop more precise intelligence for decision makers.
Basic machine literacy
Basic literacy is the ability to communicate through reading and writing. British mathematician Alan Turing’s experiment in 1950 involved three humans and a machine that communicated with one another through basic literacy functions. The object of the experiment was to see if the humans could conclude which character was the machine. Although the machine was unable to completely fool the three humans, Turing’s experiment did show that machines had the potential to exhibit basic literacy skills.
Dr. Hugh Loebner, a New York Philanthropist, designed a contest in 1990 challenging individuals to construct a machine, or bot, whose “Responses were indistinguishable from a human’s.” (The Loebner Prize—‘The first Turing Test,’ http://www.loebner.net/Prizef/loebner-prize.html ) A few contestants have developed bots with advanced basic literacy skills, but none have indistinguishably simulated human literacy and earned Loebner’s $100,000 grand prize.
You can access various bots on the internet today capable of interacting in basic conversation. Imagine a combination of a mysimon.com proprietary algorithm and a literacy proficient bot? A virtual salesperson could answer any of your questions as well as suggest various products or ideas that correlate to your criteria and interests.
The distinction between artificial intelligence and business intelligence (BI) is becoming blurred. AI will continue to produce faster, more reliable intelligence as computers become more analytically proficient. AI and BI can compliment each other by providing new techniques that affect the way businesses implement strategies.
Independent Industry Analyst Philip Russom states that “Many companies that depend on business intelligence as an important component of business relationships now need to move to the next level by implementing a 'BI network,' which is where multiple extranets are networked via web services.” (Philip Russom, “Microstrategy, Best in Business Intelligence,” http://www.microstrategy.com/Company/Analysts.asp , October 2003 .) This burgeoning relationship between AI and BI will have a significant impact on the strategies of business activities by reshaping future goals and objectives.
AI has the potential to take over a significant role in the daily lives of analysts and is continuing to shape the way in which analysts perform their jobs. AI is already capable of accomplishing parts of the abilities and skills once executed by analysts and may be able to outperform many aspects of analysts by the year 2030. AI is the engine to future business operations and will be prevalent throughout businesses of all sizes in the near future. Analysts will need to accept a new realm of intelligence activity rather than competing with it.
Background
Philip J. Annibale is a graduate student at Mercyhurst College in Erie, Pennsylvania, and is studying in the Research Intelligence Analysis Program curriculum. For his research, he chose to explore the potential of artificial intelligence, its impact on expert system development, and how it will affect the future of Intelligence Analysis. Philip’s undergraduate major is in accounting. He can be reached at Phil_annibale@hotmail.com
scip.online, number 45, December 23, 2003.
copyright Society of Competitive Intelligence Professionals, www.scip.org
[PRINTER FRIENDLY VERSION]
|
|
|
There are no letters available.
|
|
[POST]
|
|
| |