HFTP Blog
July 25, 2024

AI 101: A Primer on How It is Structured

Research>Scholars
Written by Frederike Czerlinski — Contributor

Artificial intelligence (AI) is a term that dominates conversations in many contexts. However, many people struggle to provide an explanation of what AI is and how it functions. The release of OpenAI's ChatGPT in 2022 has reignited the hype surrounding AI, but this excitement is not new. AI has been a topic of interest since 1956, when John McCarthy organized a conference in Dartmouth, New Hampshire, where scientists gathered to discuss and explore the possibilities of AI. The term "artificial intelligence" was first used at this event, which marked the beginning of a field that has continued to fascinate and develop ever since.

How is AI defined?

“Artificial intelligence is the attempt to simulate cognitive human intelligence in machines in order to use them for the benefit and advantage of humans" (Otte, 2023, p. 36).

In the absence of a universally accepted definition of AI, many authors approach the concept through the lens of human intelligence. Human intelligence itself is a subject of countless interpretations since it is defined by humans themselves, but it is often associated with logical thinking. Intelligence can be defined as "the sum of cognitive and perceptual processes of an object or subject to appropriately respond to environmental and surrounding influences" (Otte, 2023, p. 30). This includes the ability to learn from observed errors and modify one’s behavior to respond adequately in similar situations. This is similar to how AI works.

What can AI do?

AI technologies are designed and trained to interpret data, learn from it and make decisions or perform tasks that would typically require human intelligence. It can be broadly categorized into two types:

Weak AI is designed to perform a specific task. It is the type of AI we already use (e.g., email spam filters, voice assistants like Siri and Alexa, or recommendation algorithms on Netflix). These systems are highly effective within their defined boundaries but cannot operate beyond their programming.

Strong AI represents a more advanced form of AI that exceeds these abilities. It aims to replicate human consciousness and emotional intelligence. However, this remains largely theoretical and has not yet been achieved. While the effects of AI are being discussed from ethical and legal perspectives, there is yet no reason to fear AI.

How does AI learn?

AI can only show intelligent behavior if it has a solid technological foundation. While AI is a highly complex field, this section will outline the most important facts.

Machine learning (ML) is the core of many AI systems. It involves training algorithms on large datasets to learn patterns and make predictions. There are three main types of machine learning:

Supervised learning: Algorithms are trained on labeled data, meaning the input and the desired output are provided. The system learns to assign the inputs to the outputs to make accurate predictions. The training is repeated until the algorithm delivers reliable results.

Unsupervised learning: The system is given data without explicit instructions on what to do with it. It must identify patterns and relationships within the data independently.

Reinforcement learning: The algorithm learns by trial and error, receiving rewards or penalties based on its actions.

Current sources add self-supervised learning, where the algorithm generates its own labels from the input data. This allows the AI to predict data properties such as the next word in a sentence based on context, which forms the basis for systems such as ChatGPT.

Deep learning, as a subset of ML, uses neural networks with many layers (hence "deep") to analyze various factors of data. In contrast to ML, deep learning can process complex and unstructured data such as images, audio and text directly, making it a powerful tool but difficult for humans to follow.

How to use Generative AI

The most well-known type of AI is generative AI, also called “creative AI”, which uses algorithms to create new data, texts or images. Applications like ChatGPT or DALL-E can be used in both personal and professional contexts. However, it is essential to know how to use them effectively.

Here are the key elements to consider when crafting a ChatGPT prompt:

Role: Define the role or persona of the AI. For example, "You are an experienced (…)."

Context: Provide context to guide the AI’s response. "Please help me with (…)."

Example: Offer a specific example to clarify your request. "Take an example from (…)."

Tone: Specify the desired tone of the response. For example, "The tone should be (…)."

Target Audience: Identify the intended audience for the response. "My target audience is (…)."

By paying attention to these factors, you can ensure that the AI provides useful and relevant outputs.

Conclusion

The advancement of AI technology is benefiting numerous areas of society, such as medicine or autonomous driving. It can solve complex problems quickly. As AI affects our daily lives, it's important to understand how to interact with it in order to harness its full potential.

Frederike Czerlinski is a research scholar with the HFTP Middle East Research Center and student at the Emirates Academy of Hospitality Management in Dubai, UAE.

References

  • Beck, S. (2020). Künstliche Intelligenz – ethische und rechtliche Herausforderungen. In: K. Mainzer, (Hrsg.) Philosophisches Handbuch Künstliche Intelligenz. Springer Reference Geisteswissenschaften. Springer VS, Wiesbaden.

  • Dahm, M. H., Zehnder, V. (2023). Grundlagen der KI. In M. H. Dahm & V. Zender (Hrsg), Moderne Personalführung mit Künstlicher Intelligenz. Chancen und Risiken. Springer Gabler, Wiesbaden.

  • Eichstädt, T., Spieker, S. (2024) Künstliche Intelligenz. In T. Eichstädt & S. Spieker (Hrsg.), 52 Stunden Informatik. Was jeder über Informatik wissen sollte. (2. Aufl., S. 373-382) Springer Vieweg Wiesbaden.

  • Günther, J.-P. (2016). Robotik und rechtliche Verantwortung. Eine Untersuchung der Benutzer- und Herstellerhaftung, Herbert Utz Verlag.

  • Hamet, P., & Tremblay, J. (2017). Artificial intelligence in medicine. Metabolism: clinical and experimental.

  • Haenlein, M., & Kaplan, A. (2019). A Brief History of Artificial Intelligence: On the Past, Present, and Future of Artificial Intelligence. California Management Review.

  • https://www.instagram.com/renate.gpt/

  • Krauss, P. (2023). Wie lernt Künstliche Intelligenz. In P. Krauss (Hrsg.), Künstliche Intelligenz und Hirnforschung (S. 125-238). Springer, Berlin, Heidelberg.

  • Kreutzer, R. T. (2023). Was versteht man unter Künstlicher Intelligenz und wie kann man sie nutzen?. In R. T. Kreutzer (Hrsg.), Künstliche Intelligenz verstehen. Grundlagen – Use-Cases – unternehmenseigene KI-Journey (S. 1-114). Springer Gabler Wiesbaden.

  • Lu, H., Li, Y., Chen, M., Kim, H., & Serikawa, S. (2017). Brain Intelligence: Go beyond Artificial Intelligence. Mobile Networks and Applications.

  • Mueller, J. P., Massaron, L. (2017) Maschinelles Lernen mit Python und R für Dummies. (1. Aufl.) Wiley-VCH, Weinheim.

  • Otte, R. (2023). Künstliche Intelligenz für Dummies. (2. Aufl.) Wiley-VCH, Weinheim.

  • Wittpahl, V. (2019), Künstliche Intelligenz. Technologie – Anwendung – Gesellschaft, Springer Vieweg Berlin, Heidelberg.

  • Image: https://www.gabo.de/kuenstliche-intelligenz-vs-machine-learning-definition-und-abgrenzung/

2024 research Middle East Research Center scholar AI definitions generative AI Artificial Intelligence machine learning deep learning