In recent months, particularly following the release of ChatGPT, there has been an unprecedented surge in interest surrounding artificial intelligence (AI). This heightened attention spans across a multitude of sectors, including business enterprises, technology companies, venture capital firms, universities, governments, media outlets, and more. As the interest in AI is intensifying, some companies have even rebranded their existing software solutions as “AI” products, a phenomenon often referred to as “AI washing.” Furthermore, there is also a growing sense of “FOMO” (Fear of Missing Out) among corporations regarding AI adoption.

So, what exactly is AI? In simple words, AI refers to the simulation of human intelligence in machines that are programmed to think, learn, and make decisions. An AI solution has five key building blocks.

1. Data: Data means numbers, characters, images, audio, video, symbols, or any repository on which operations can be performed by a computer.
2. Algorithm: An algorithm is a sequence of calculations and rules used to solve a problem using data that is optimized in terms of time and space.
3. Model: A model is a combination of data and algorithms used to generate the response. Once you have a model, you can constantly provide it with new data and algorithms and continue its refinement. 
4. Response: The responses are the results or outputs from the models. The outputs are based on the specific objectives that could be related to automating tasks, providing insights, aiding decision-making processes, and more.
5. Ethics: Ethics refers to the moral principles and guidelines governing the collection, processing, analysis, interpretation, and application of data and insights in AI. Ethical considerations are crucial in ensuring that data-driven outputs contribute to the positive social, economic, and environmental impacts of the organization and the community.

However, the term “AI” often causes confusion due to its broad and sometimes vague usage. A “true AI” system, composed of an agent that performs the task in the environment, has three key characteristics:

1. Learning: The ability to learn from data and improve over time without explicit programming.
2. Adaptability: The capability to adapt to new situations and use cases beyond their initial or original purpose. An AI system should have the capability to reason or think and address the objectives through logical deduction.
3. Autonomy: The AI system should perform tasks independently with minimal or even zero human intervention.

Practically, AI can work in any situation where one can derive patterns from data and formulate rules for processing. In other words, AI systems perform poorly in unpredictable and unstructured environments where there is a lack of clear objectives, quality data, and predefined rules. While AI can analyze vast amounts of data, identify patterns, and derive rules, it cannot generate truly new hypotheses. True innovation often requires intuition and a solid understanding of broader innovation principles and practices. Last but not least, AI can struggle with ethical dilemmas and making decisions that require moral reasoning, empathy, and understanding of human culture and values.

So, what are the real use cases of AI? Where are the three AI characteristics discussed above applied or used? An autonomous vehicle is a classic example of an AI solution that applies the three AI characteristics, i.e., learning, reasoning, and decision-making, in real time, to create a vehicle capable of driving without human intervention. Waymo’s autonomous vehicles are equipped with a suite of sensors, LiDAR and radar capabilities, and high-definition cameras to collect massive amounts of data about the vehicle’s navigation and its surroundings. Advanced machine learning algorithms are used to process and interpret this data. These models are constantly trained on vast datasets, allowing the car to recognize and categorize new objects and situations, predict the actions of other road users, and make real-time driving decisions to ensure safe and efficient driving.

Another use case where the three characteristics of AI – learning, reasoning, and decision-making – are used is in writing a book. AI systems like ChatGPT are trained on extensive datasets such as books, articles, and other types of content. This training enables the AI to understand language patterns, narrative structures, and stylistic elements. AI can gain knowledge to understand the components of a story and reason about the characters’ motivations to create engaging plots. As the story progresses, the AI can make decisions about plot twists and character actions. Here is an example of how ChatGPT, which uses the GPT-4 architecture, wrote a story about me (even though some of the response is inaccurate).

In contrast, certain technologies may be sophisticated but do not qualify as AI if they lack the characteristics of learning, reasoning, and autonomous decision-making. For example, one can use robotic process automation (RPA) tools to automate repetitive tasks such as data entry, data cleaning, data validation, processing, and more as per defined rules and thresholds. RPA is not AI, as it lacks the learning and adaptability characteristics (characteristics No. 1 and 2) of a true AI system. RPA systems make decisions and implement them based on a bunch of if-then rules resulting in clear and deterministic outcomes.

Similarly, chatbots are not AI. Chatbots use scripted responses and pattern matching to engage with users. Voice assistants like Apple Siri, Amazon Alexa, and Google Assistant are also not AI. They use a combination of speech recognition, NLP, and predefined responses. A linear regression model to predict sales based on historical data, lacks the adaptability and complex decision-making capabilities typically associated with AI. Regression models primarily help in forecasting. But all these “AI tools” have some degree of the three fundamental characteristics of AI, and much of their functionality relies on scripted responses and simple data processing based on predefined rules. 

AI today has become more than just a technological advancement. It symbolizes productivity, innovation, risk, and opportunity. What sets AI apart is its ability to learn, adapt, and make decisions based on data, often mimicking aspects of human intelligence. Systems or processes that do not incorporate these characteristics do not fall under the umbrella of AI. Overall, while the enthusiasm surrounding AI is palpable, it would be prudent to approach its adoption with a balanced and responsible understanding of its capabilities and limitations so that the true value of AI is realized.