Artificial Intelligence: Human Factor Failure, the core values (CVI) approach to comparing IA versus AI systems. #4

Full Publication and Speaking Tour 2020, follow #FOCUS2020 for details. #AI

SEE ARTICLES: #1, #2, #3, #4, #5, UX#1, UX#2

Artificial Intelligence: Human Factor Failure

The #FOCUS for this series is to start the discussion around the ethical framework for development of AI. The primary problem with the artificial intelligence (AI) industry is the over reliance on computer-processing bias. Research is occurring in computer-process vacuum. To address this situation in more depth, this article section focused on the role of human decision making from two vantages points: psychology of heuristics, and organizational-development though an independently validated tool, the CVI. Arguably, if humans do not adapt, they will be cornered out of one of the fastest growing labor economic markets in modern history.

Take for example the MIT system on human factors, the entire staff is made up of computer scientists. Not a single sociologist, psychologist, coach, business professor, nor any connection to the world other than computer science - Yet, the program is geared toward teaching HUMAN principles? This is akin to asking McDonald's to produce Toyota's. We would not allow such to be taught, so why are computer scientists teaching advanced courses on human thinking and human factors with no experts in the field? This program is not alone, in fact virtually every university certificate course and program is over leveraged in the computer science realm. Then again - maybe the hyper #FOCUS on ONE human thinking type will limit the evolution of AI for the better? This is WHY the ethical framework directing the education, creation, research and monitoring of AI should happen NOW, and not after the fact.

The Core Values Index (CVI)

The Core Value Index (CVI) is an organizational development tool. The tool is based on four factors of heuristic decision making. These factors pertaining to specific traits and states that attract individuals to specific types of job functions. The CVI provides scores in each of four domains: innovator, merchant, builder, and banker. While an individual will have scores in all areas, there is often one primary with a distal secondary domain. Each domain has an entire method fo establishing thinking, learning, communication systems relative to an individual's score. The CVI has a extremely high reliability score                

The Core Values Index? (CVI; formerly PVI) has been an established psychometric tool created and provided by Taylor Protocols. Taylor Protocols has tested several thousand individuals using the CVI assessment over the past decade. Seattle Research Partners, Inc. was provided the raw database for test and re-test participants having taken the CVI at uncontrolled intervals. Over the past three years; Seattle Research Partners has had three consultants complete the CVI. The CVI assessment and raw score entries were located and determined the raw data access represented true and accurate information captured by Taylor Protocols over the past ten years. The data was compiled and prepared for statistical analysis by Seattle Research Partners, Inc. consultants.

 The objective for this specific report was to establish and independent reliability score for the CVI using best practice methodologies. This is the first of several intended independent research studies to understand, validate, and report on the Core Values Index? assessment as well as the Taylor Protocols’ methodology for best practice use of the CVI. To our knowledge the CVI assessment has remained unaltered in its construct in capturing the innate energies of humans for over a decade. Drawing from thousands of individual scores, a reliability study was conducted based on available test-retest results of random individual taking the CVI assessment.

Findings: Reliability Test

Seattle Research Partners was asked to make a study of the reliability of the Core Values Index (CVI), a psychometric instrument that has been used by Taylor Protocols for several years. Taylor Protocols supplied test-retest results from past clients, from 2002 to 2013. The sample size used was n= 711. The method used was test-retest, with regression constant set to zero. Individuals were re-tested at various intervals, ranging from just a few months to over ten years. Seattle Research Partners, Inc. cross-checked reliability results by using two separate SRP researchers. Both professionals hold a doctorate degree; one holding a doctorate in statistical analysis and the other in psychology.

Regression Statistics (reports by SRP, Inc. 1999|2014; RSolutions, 2018)

               Findings based on Multiple R value are presented below.

Multiple R         0.97697 R-value (“Correlation”) represents the relationship between the

test-retest scores. A relationship correlation of 98% is extremely high and demonstrates the reliability of the instrument.

R Square             0.95446

Adj. R Square    0.95399

Standard Error  4.058

Observations    2133

P-Value                               0

The low p-value and the high correlation both indicate that first test scores are a highly significant indicator of retest scores. We independently conclude that the CVI is a reliable instrument; reliability remains constant through a three year basis based on random sampling.

Summary

The low p-value and the high correlation both indicate that first test scores are a highly significant indicator of retest scores. We independently conclude that the CVI is a reliable instrument. We are not able to comment at this time on validity or bias in the CVI, having only examined the summary data from each test. Bias could be examined by a study that included responses to each of the 72 questions. Validity would require additional research tests and measures currently in process. However, validity and bias were addressed in an independent research report in 1999 in addition to reliability. At that time the CVI data, formerly the PVI, was independently analyzed for both bias and validity. In this report, there was no bias found and validity was determined to be accurate.

The CVI can be united within human decision-making models and applied to understanding the relationship of IA to AI. This is addressed in the next segment.

Human Decision-Making Modalities: Core Values Index (CVI) with heuristic factors

The significance of both related to AI and technology is rather clear when allowed time to examine the relationship: as technology increases in development, design, and innovative interactive consumer uses, it requires the ability to standardize a framework for customized individual interaction (also known as an applied-learning platform). Such platforms have enabled AI to learn possible wants, interests, habits and information to help predict and provide useful information based on data captured for each user.

This process creates a four factors of User Experience research modeling. This is a multi dimensional model, so keep that in mind as you read as each of the XY axis and vertices converge at a “FOCUS point”. Think of a three-dimensional + (top/bottom, sides) for each of these factors that connect at a single common core single focus point:

• Artificial Intelligence (AI) - machine learning, deep-learning, and responsive technology
• Individual Authenticity (IA) - reviewing and using heuristic processes to make quick decisions based on a framework of human thinking patterns
• artificial-intuition (ai, lower case) - machines ability to predict useful information for unique users based on device and data storage and prior experience records
• intuitive-assessment (ia, lower case) - man's ability to predict useful information based on human factor experience

The ultimate goal of technological utopia: “The Organic Machine” a linking of individuals and machine in a manner that the machine enhances the human existence without usurping control or experience. The ultimate disaster - a thing of sci-fi movies, but not all that unrealistic when man battles IA-AI linked self-governing machine. To achieve the positive outcome, “The organic-machine” it is important to understand the four-modalities of thinking represented by each factor (AI, IA, ai, ia) of human thinking and experience.

There are four modalities of heuristic decision-making within applied-research (SLA, 2015; RAND, 2018). These are:

• global-spatial,
• Process-coding,
• Process-outlining,
• Global-Spiraling.

A global-spatial individual thinks of the self in relationship to filing separate buckets. For practical human factors description, consider the core values index (CVI) description of a “merchant” (for an in depth an independent article On Core Values, click here). The “merchant” acts to sell others specific “products” or “topics” of interest. They can speak to a variety of topics, much like a merchant will have a general knowledge of all items within their stores. The merchant knows where to go to grab an item based on where it is "stored" (similar to a concept of nodule-storage in both heuristic psychology and machine learning/deep learning constructs). Within this nodule-storage, a merchant will know the general characteristics comparing and contrasting similar “product” or topic constructs. The merchant will rely on AI to identify the consumer experience and line shelves with corresponding products or interests. Sometimes “teasers” are meant to help spur more purchases, in the same manner AI works to recognize consumer and user needs and respond.

A process-coding individual, however, is best defined by the CVI construct of being a builder. Builders do. Builders do based on a plan. Give the builder a plan and empower the builder to create, cultivate, and generate in depth relationships, programs, and concepts. They go direct to those who have building opportunities; they are not concerned with the plannings, the tinkering, the dreams or the wants, just the practical application and output. In machine-learning and data-science, these individuals think along linear lines. They are, after all, coders.

A process-outlining, on the other hand, is like a CVI banker. They are interested in the data, they are interested in the structure of the data and uniform principles for connecting the data. They want to see very clear rules establishing when, where, how, and how frequently to count data, dollars, outcomes. Process-outlining takes time to work down and operated very similar to computer-processing. Thus process-coders (builders) and process-outliners (e.g., data-scientists) are often characteristics of people who are drawn to banking, engineering, municipal planning, accounting, data science, research science, sales, and operations/process building. These individuals rely in Individual Authentication (IA) to inform and instruct the creation of AI systems through blueprint planning. They also rely on intuitive assessment for building codes because there often is more than one method to obtain an outcome (the challenge being a task of efficiency for the builder-types)

Global-spiral thinkers per the CVI are innovators. They are the people who see connections between unrelated objects. The are comfortable in the world of lower-case ia! Intuitive authentication. They can fill the gap between stimulus and response with what-if creations, alternates, and new world visions. While they may be defined as disruptors, they are only in as much as dreams are disruptive. After all, they are not inspired by pedaling their services, they are interested but only at a high-level to the needed accounting and coding of data and dollars, and they are polar opposite the builder.

A spiral thinker, unlike a global-spatial thinker, is always mixing and matching by aligning possible “connections” between concepts that are similar but otherwise unrelated. Imagine going to the “merchants store” where everything is arranged based on needs for the day by price point. Low cost, low quality; high-cost, high-quality. Walk the isle and find the different items you might need - breakfast to dinner. Therefore, the spiral-thinking individual will rely more on intuition more than their similar counterpart, the merchant, that relies on the data-driven information of product alignment, shelf/space efficiency provided through AI analytics.

Analysis:

CVI and learning these thinking patterns... GET THE FULL PUBLICATION 2020!

Discussion:

Implications for human learning when computer science allows a forced-fit over human factors.

• Is this healthy?
• Is this ethical?
• What obligations do higher education providers have to ensuring the health and future of humans who think in diverse patterns?
• Where is the initial inclusion in the development of AI and its influence over IA?

GET THE COMPLETE PUBLICATION #FOCUS2020

? 2019 | ORCHARD-PRESS, a division of RSolution Publishers & in cooperation with LINKED-IN, Strategic Learning Alliance, and Saber-Mountain Press.

All rights reserved. No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the publisher, except in the case of brief quotations embodied in critical reviews and certain other noncommercial uses permitted by copyright law with specific reference and citation. ANY USE OF THIS CONTENT WITHOUT PERMISSION OF THE AUTHOR, PUBLISHER, OR ASSIGNS IS ILLEGAL.