Join our book community on Discord

https://packt.link/EarlyAccessCommunity

The previous chapter introduced us to the world of analytics, including key analytic concepts and the Analytics Maturity Index. But AI is too much of a game-changer not to give it its own chapter. AI is different from any other analytics technology we have seen. Unlike traditional analytics that seek to optimize operational use cases, AI seeks to learn. And AI can continuously learn and adapt at speeds billions if not trillions of times faster than humans with minimal human intervention, without the proper precautions and AI model training, that can lead to disastrous unintended consequences.So, let’s use this chapter to deep dive into how AI works and the vital role that every person needs to play in defining the AI utility function that guides how the AI model works to achieve user-defined outcomes.This chapter will cover the following topics:

• How does AI work?
• What constitutes a healthy AI utility function...

How does AI work?

An AI model is a set of algorithms that seeks to optimize decisions and actions by mapping probabilistic outcomes to a utility value within a constantly changing environment…with minimal human intervention.

Utility value describes the subjective value the AI model assigns to different outcomes based on its programmed goals and objectives. For example, an AI designed to maximize profit for a company may assign a higher utility value to decisions that result in increased revenue or decreased expenses. In comparison, an AI designed to improve healthcare outcomes may assign a higher utility value to decisions that lead to improved patient outcomes and reduced healthcare costs.

This set of algorithms can learn trends, patterns, and relationships buried in data and make predictions based on those trends, patterns, and relationships to make decisions and act on them. The AI model is trained on a dataset consisting of variables and metrics (input data) that corresponds...

What constitutes a healthy AI utility function?

The AI utility function is a set of factors (variables and metrics) with associated weights that map outcomes to utility values to guide an AI model's decision-making process. It measures decision effectiveness and continuously learns and adapts to improve performance. It is a mathematical expression that maps the input space (e.g., the possible actions an AI model can take) against a set of output values representing the operating environment’s preferences or goals. The weights assigned to different variables and metrics determine the relative importance of those variables and metrics in the AI model’s decision-making process. By maximizing the expected utility of its decisions, an AI model can make choices more likely to achieve the desired operational outcomes.An AI model will continuously seek to optimize its AI utility function as the AI model interacts with its operating environment. The AI model provides positive...

How to optimize AI-based learning systems?

Next, we want to expand your role in driving the AI conversation by explaining the importance of first understanding or determining user intent and then discussing the importance of defining conflicting variables and metrics and how multi-objective optimization will guide the trade-off decisions and actions that the AI model must make. Let's understand this with an example!Whether you use Google Maps, Apple Maps, or Waze (also owned by Google), these AI-infused apps are fantastic at getting you from Point A to Point B in the shortest time. They give you step-by-step directions and can update those directions based on current information (e.g., traffic accidents, weather, potholes, special events) that pop up during your trip.Let’s consider the different variables and metrics that we would want to include in the AI utility function to help the GPS optimize our journey. Two classes of variables and metrics to include in the AI utility...

Summary

AI is the ultimate goal of an organization’s analytics maturity. AI builds upon the analytic maturity levels of Operational Reporting, Insights and Foresights, and Augmented Human Intelligence to create Autonomous Analytics that power the products, services, policies, and processes that can continuously learn and adapt to the changing environment in which it operates. This continuous learning and adapting of the AI-powered products, services, policies, and processes can happen with minimal human intervention, which makes it unlike any technology that we have ever seen.To become a Citizen of Data Science, everyone must understand how AI and the AI utility function collaborate and hopefully this chapter has helped with that. That’s fundamental to participating in the AI conversation and debate. But everyone also needs to understand their role in ensuring the development of a heathy AI utility function that guides the AI model to deliver meaningful, relevant, responsible...

Critical thinking in decision making

Critical thinking is the rational and objective analysis, exploration, and evaluation of an issue or subject to form a viable and justifiable judgment. However, to truly understand critical thinking, I believe that one must first understand objectivity.

Objectivity, at its core, is the foundation for making intelligent and well-informed decisions. When engaging in the decision-making process, it is crucial to approach the process with an open mind, free from preconceived notions. If you already have a pre-determined decision in mind, you run the risk of selectively seeking data that confirms your position and disregarding information that contradicts your beliefs. Consequently, prioritizing objectivity and consciously avoiding personal biases are essential for mastering critical thinking and thus informed decision-making.

As discussed in Chapter 1, organizations with various missions gather your personal data to influence your decisions...

Summary

Everything about AI and data literacy can be applied to help us make more rational and informed decisions. Decisions represent the crucial juncture where theory and practical application merge, ultimately benefiting society as a whole. By harnessing the power of AI and understanding how to effectively interpret and utilize data, we empower ourselves and society to make choices that are grounded in reason and knowledge, leading to positive outcomes for all.

And while not every decision will require a formal decision matrix process, you’d be surprised how quickly one can develop informed decision-making as a muscle memory. And in a world where organizations are trying to influence your behaviors, beliefs, and actions through half-truths, white lies, fake news, and alternative facts, informed decision-making is an invaluable skill.

In the next chapter, we will delve into the fundamental statistical concepts that are essential for individuals and society to develop...

References

1. Sensitivity Analysis: https://www.investopedia.com/terms/s/sensitivityanalysis.asp
2. Monte Carlo Technique: https://www.investopedia.com/terms/m/montecarlosimulation.asp
3. Toronto Star. When U.S. air force discovered the flaw of averages: https://www.thestar.com/news/insight/2016/01/16/when-us-air-force-discovered-the-flaw-of-averages.html
4. Harvard Ed Magazine. Beyond Average by Todd Ross: https://www.gse.harvard.edu/news/ed/15/08/beyond-average

Join our book’s Discord space

Join our Discord community to meet like-minded people and learn alongside more than 4000 people at:

https://packt.link/data