Why AI will not replace human intuition but will optimize it
Intuition is the oldest form of predictability, and therefore, it is the resource most used by decision-makers who today face the challenge of doing their own thing in the face of a huge amount of data. This, which could be a great opportunity to make business profitable and have control of valuable information, is, for many, a bottleneck that ends up in the management of errors rather than successes. Is Artificial Intelligence the lifesaver of companies?
The influence of intuition, experience and data in business decision making is changing radically. We are facing a period of transition where collaboration between people, Artificial Intelligence and the predictability it brings, is being reconfigured to open the doors to much more timely dimensions.
Until a few decades ago, to manage an important but limited volume of data, it was enough to have a good method. Later, with the technological evolution, companies were able to count on systems that allow their centralization and draw certain conclusions from the information already processed, which is why Business Intelligence, with its traditional analytical reports, became absolutely vital for companies of a certain size.
But in this decade, with the advent of Big Data, it is not possible to work with methodologies or tools created for a limited volume of data, only. Not even machine learning is enough for what is to come.
It is at this point that AI becomes central: when the volumes, variety and sophistication of the data generated at the organizational and market levels are overwhelming, both for the human mind and for the tools we have at our fingertips.
AI is on its way to becoming indispensable. In my humble opinion, together with the automation of processes and robotization, we will have a greater impact than resources such as electricity or the Internet have had on humanity.
Of course, the existence of true artificial intelligence is essential in this technological context, so that decision making becomes real. Why? Because, unfortunately, and although it is very common to hear about AI related to multiple tools and platforms, in most cases it is machine learning or automatic learning but not Artificial Intelligence itself.
In other words, the system is created to classify according to possible contexts that have already been incorporated and it does so through labels that have previously “taught” it what something actually is.
True AI must be able to perform three operations: learning, reasoning and predicting on its own.
As Yann LeCun, head of AI at Meta, says, “One of the most important challenges in AI today is to design learning paradigms and architectures that allow machines to learn models of the world in a self-supervised way and then use those models to predict, reason and plan”.
As soon as we see this topic appear, some arguments arise as a kind of defense, animated by fear or threat. On the one hand, the loss of analytical capacity and agility that comes with dependence on technological tools that “think” for us.
We no longer know how to get anywhere without Waze. These kinds of tools that read volumes of data and make a simple prediction of what is the best road for the current distribution of vehicles on the streets, somehow make us more lazy when faced with the task of thinking of alternative routes, for example.
But the biggest threat it seems to generate is the loss of power implied by delegating a part of decision making to the workings of AI, an entity that most people barely understand and that has a processing capacity unmatched by a human being.
In the traditional decision-making process, we could say that only 30% of what influences a decision is the information we can process, and the rest is intuition or experience, but this is no longer sustainable.
According to a recent Gartner survey of senior executives of leading companies, 65% agree that the decisions they make are more complex than they were two years ago and need to be explained or justified.
Moreover, the post-pandemic scenario is characterized by two major phenomena that have a direct impact on decision making for companies, that is:
- The acceleration of digitization in 2020 and 2021 that leaves us today with a data production that we have not yet fully sized
- The context of global uncertainty on multiple levels (economic, social, work, business, technological, etc.) that still remains.
It is clear that both phenomena have become part of our reality, where the first, the exponential generation of data, improves decision making if it were possible to manage and benefit from them. But the second phenomenon, uncertainty, makes the task more complex for the players who must establish or plan strategies for the future.
This is why we have to rethink the tools we have and will have in the future to be able to face these contexts and make strategic decisions in an agile way
Having optimized and digested information as input has become essential for decision making, there is no doubt about that, but if we combine this with the possibility of adding new patterns and simulated scenarios that generate a new probabilistic dimension, we could approach a much higher level of certainty in our decisions about the goals set by each organization.
According to a survey on AI in business conducted by PWC in March of this year, the leaders questioned recognize that the main application value of AI is in the automation of processes, and the second one is in the improvement of the decision-making process.
In those companies that are already considering how to handle the interaction between AI results and humans in the business decision process, there seem to be two opposing roads: one that proposes to fully delegate decision making to the results of AI and predictive tools, and another, where there are those who do not trust them and prefer to ignore them.
Personally, I believe that in the current state of affairs, neither one nor the other is wise. Even in companies that already incorporate technology to automate processes, they are not considering delegating decision making, but rather improving the process through interaction with AI results.
This is an important difference because it gives us the guideline that today it is not possible to imagine a sufficient state of development in which the predictive models that can be provided by today’s AI can replace human judgment to any extent, imbued with knowledge, experience and common sense that is still impossible to transmit to an algorithm.
The reality is that the technology available to companies today is in a very early stage where there is still a huge opportunity for automation.
By incorporating machine learning tools, it is possible to “teach” a system what humans do repetitively, but we have not yet reached the point where it can reason for itself and predict scenarios in a way that surpasses the good sense of a manager with years of experience.
To be ready for the potential that AI will bring to companies, they must incorporate automation tools today and start down the path of interaction between the human factor in decision making and the capabilities that systems and algorithms bring to the table.
Only then will we be able to digest this transition towards a balance between intuition, experience and the predictability that AI can bring in a world in which strategic decision making is becoming increasingly complex.
The world is moving on, technology is moving ahead, decision making scenarios will become more and more complex and Artificial Intelligence, as long as all this continues to develop, will be a necessity and not a privilege reserved for large companies.
Original column published in https://www.infobae.com/opinion/2022/12/05/por-que-la-ia-no-reemplazara-a-la-intuicion-humana-pero-si-la-optimizara/