Why artificial intelligence? "I know quality when I see it!"
12th January 2021
By John N. Moye, author of A Machine Learning, Artificial Intelligence Approach to Institutional Effectiveness in Higher EducationArtificial Intelligence (AI) is a product of the analysis, modelling, and sensemaking of organisational data into an interconnected and aligned intelligence. It is produced by the systematic measurement, aggregation, and modeling of authentic, valid, and reliable organisational "performance" data.
As a result, the actual performance of the institution is exposed and used to understand and adjust organisational performance to reach the desired state. AI realises the goal of many organisations to create an "evidence-based", "data-driven" process of organisational management and improvement as a complex phenomenological construct.
AI does not eliminate or replace human intelligence and decision-making within an organisation. Instead, it provides an objective and accurate description of the actual performance of the organisation and its systems by creating authentic performance models based on the analysis of reliable and valid measurements of functional performance. The structure of this AI is determined by the interconnections and interdependencies that emerge from the analyses of the performance data, which describe the behaviour and performance of the organisation without assumptions or predictions. Therefore, the AI generates its conceptual framework and generates a complex description of the performance of organisational systems as measured directly by the performance of the functional systems.
The intended performance of the organisation is framed by established models of organisational effectiveness. These effectiveness models structure the intended performance data and create a conceptual framework for the comparison of actual and intended performance information. The models of organisational effectiveness are chosen and determined by the human intelligence in the organisation. The selection of the effectiveness model articulates the structure of intended performance.
Organisational performance is often assumed to be a simple linear process of cause and effect, which is measured and demonstrated by individual data points (goal attainment and strategic contingencies approach). However, the structure of most 21st century organisations is more complex than these mechanical models, especially organisations that serve humans. These organisations require organic models (strategic constituencies organisation approach) to reflect the intended performance. The organic models frame performance as a complex, dynamic, and non-linear phenomenon to reflect the actual structure of the processes an organic organisation employs to deliver its mission (systems and competing values organisation approaches). However, all organisations employ all five models at some level in their performance a complex mission and the interdependencies between these models generate their own artificial intelligence generated by the human sensemaking.
There are five different approaches to organisational performance that can be used to align and differentiate intended performance data by the processes of organisational performance: the goal attainment, the strategic contingencies, the strategic constituencies, the systems model, and the competing-values approach. Each of these approaches assumes different outcomes and processes and organise performance data to align with the characteristics of intended organisational functions to produce an authentic description of intended organisational performance.
Table 1 differentiates the characteristics of these models on four performance assumptions to reveal the conceptual framework that underpins them. Column 1 identifies the organisational approach to effectiveness, column 2 specifies the type of process model assumed by the approach, column 3 identifies the type of result the approach intends to deliver, column 4 describes the assumptions regarding the outcome of the approach, and column 5 presents the model of thinking assumed by the approach.
Table 1: Differentiation of Performance Models
The data analysis methods used to uncover the interrelationships and interdependencies within organisational data are codified in the field of Machine Learning (ML). The data modelling techniques of ML provide a plethora of objective methods to create authentic AI from organisational performance data. The goal of most ML strategies is to uncover the information contained within a dataset collected from a system or organisation (machine) and describe the performance of that "machine" by modelling those data.
While organisations are composed of and created by humans, organisation effectiveness studies the performance of the organisation, not the individuals within it. An effective organisation employs systems that produce the desired result and preclude the ability of a human to short-circuit that system. Ineffective organisations design systems that produce different results than those desired, regardless of the efforts of the humans within the organisation.
The ML strategies and processes uncover the mathematical interconnections and interdependencies (structure) within the data. These relationships reveal the structure that is used to transform stochastic data into meaningful "intelligence". Figure 1 depicts the ML processes that identify the structure within the data.
Figure 1: ML Analytical Strategies to Reveal the Structure of AI
The models of actual performance populate the five approaches to organisational effectiveness to align the actual performance data with the intentions of the organisation. The result is a transformation of the actual performance in the AI into models that authentically describe the intended performance of the organisation.
The resulting AI is input to a sensemaking process conducted with and by the human intelligence of the organisation to identify plausible (not absolute) meaning and causation. In this process, the human intelligence is tasked with "looking beyond the data" to synthesise and discover plausible meaning (sensemaking). In summary, the actual performance of the institution is revealed by the ML procedures, modelled into AI of actual performance, framed by the organisational effectiveness approaches, and used to inform organisational performance and improvement decisions. Figure 2 provides an overview of this systematic approach to creating AI from the raw data.
Figure 2: Overview of AI Process for Organisational Effectiveness
Collectively, these methods and procedures constitute a systematic, evidence-based, data-driven approach to the creation of authentic, credible, and trustworthy organisational performance intelligence. It allows others to "know the quality of an organisation when they see it".