Decision making is a central aspect of managerial and organizational life, a process on which generative AI will be able to impact in even significant ways. According to a recent McKinsey study, The Economic Potential of Generative AI: The Next Productivity Frontier, AI is becoming increasingly integrated into business processes, particularly in the decision-making domain, where it is set to play an increasing role in supporting decision makers. In this context, it appears important to understand how people engaged in decision-making and AI can enter into dialogue within organizations.
We propose to take a short journey through the dimensions at play: the individual dimension, the organizational scenario, and generativ AI
The Complexity of Individual Choice Between Rationality, Experience, and Emotion
At the individual level, when we are called upon to make decisions when faced with a problem, we begin a process of gathering and integrating a multitude of information, identifying strategies and tactics best suited to make effective choices. Decision making is a process composed of several stages: evaluation and establishment of preference among different options; selection and execution of an action; and evaluation of the consequences of the choice made, which depends on the difference between the consequences expected versus those experienced (Simon 1960; Brim 1962; Dewey 1978).
We are not always able to develop the process in a planned way through the adoption of explicit strategies based on comparisons and assumptions. In contrast to early normative theories (Bernoulli, 1738; Von Neumann and Morgenstern, 1943), which described a rational decision maker, we now know that humans often make decisions automatically, without full awareness of all options.
Research shows (Klein, 1992; Mc Menamin, 1992) that choosing in complex, multifaceted scenarios, such as our organizations may be in which one is called upon to make choices in a short time, under a condition of uncertainty, with goals that are not always clearly defined, in rapidly changing situations, activates in people the use of experience in the form of a repertoire of action patterns specific to a certain situation. These patterns quickly highlight the most important cues, offer expectations about what is likely to happen, identify possible goals, and suggest typical reactions to be taken in that particular type of situation.
In this way individuals can make successful decisions very quickly by combining intuition with conscious, deliberate, analytical analysis(Recognition Primed Decision).
Typical difficulties in decision making often result not from lack of information but from our emotions and projections. Emotions guide us in acquiring information, influencing evaluation and ultimately guiding choice. Although evaluating each option using a rational method might seem the best approach, this would require often untenable time, so decision makers unwittingly rely on heuristics and intuitive strategies. Indeed, human rationality is limited, subject to cognitive biases, which steer choices in favor of a quick decision, rather than a perfect one.
Considering the above and the fact that organizations reflect the behaviors of the individuals in them, let us now shift the focus to the organizational level.
Decision-making within the organization: quick decisions and continuous experimentation
If, at the individual level, decision making presents so many planes to consider in order to understand its nature, of no less complexity is deciding in organizational systems.
Every day, inside our organizations, we are called upon to make so many decisions: from the most routine ones to those that are unplanned and have a high impact on our work and our stakeholders/colleagues/collaborators. These decisions are often made under conditions of uncertainty, must be made quickly, and therefore it is not always possible to have all the necessary information available. For these reasons, it is important to adopt, where possible, a collaborative, iterative and transparent decision making approach that does not rely solely on hierarchy. An approach that values experimentation and learning from mistakes, seen as an integral part of the experience, in which feedback is central, to improve the quality of decisions, and also the different skills present within teams are used, thus helping to improve both the quality and speed of decision making.
What dialogue between Human and generative AI?
If this is the scenario in which we move - made up of individual dimensions, team dimensions, and organizational logics - how can AI fit into this complexity? How can AI fit into such an articulated, composite and complex scenario?
Generative AI can be of great support in predictive processes, but it lacks the judgment that distinguishes human intelligence, which is based on intuition, experience, emotions and relational influences. In complex situations, AI makes it possible to identify correlations between data and explore new opportunities for resolution, leveraging Deep Learning to generate new knowledge quickly. Conversely, when an organization is faced with uncertain and ambiguous situations for which no past data or evidence is available, the intuition, imagination and creativity of the human decision maker are the most effective means of identifying the best solution. In these cases, AI may not be adequate because it needs clearly delineated boundaries that define a specific area in which to process the available data and possibly create new ones.
In contrast, human decision making can be untethered from purely rational thinking and can be based on intuition and emotions.
Machines and people are ontologically different, machine-generated outputs, based on mathematical and static structures, can give rise to surprising and unexpected results for people, who, basing their decisions not only on logic and rationality, but also on emotions and experience, would probably have arrived at a different output. To the Human then the task of evaluating the results provided by the machine and deciding whether and how to use them being aware of the characteristics of AI in decision making.
Hybrid decision making
Human and AI thus have different characteristics, and in decision-making processes they can act simultaneously, each focusing on the part of the process most akin to its characteristics.
This interaction between human and artificial capabilities, we know is not without risks (data security, dependence on the machine, disappearance of certain professional skills...), at the same time not taking advantage of the opportunities that AI offers in terms of enhancing our abilities would be counterproductive. When the organization implements AI in its organizational processes, it must, therefore, not only care about their design and structure, but pay attention to the knowledge, awareness and training of people toward the potential and limitations of AI in order to make the most of its advantages without becoming its passive and uncritical employees or users.
#AI #decisionMaking #bias #rationality #experience #emotions #algorithmic
Want to learn more about it? We suggest some readings:
- Why human intelligence still beats algorithms, by Gerd Gigerenzer
Do you want to create the conditions to accompany people toward sustainable and productive use of AI in your company? Contact us, we can do something for you!