Understanding the Breadth of Artificial Intelligence Beyond Generative AI

In the current AI landscape, terms like ChatGPT and Copilot dominate the conversation, leading many to equate AI solely with generative AI. However, AI encompasses various technologies, offering unique business value opportunities and competitive advantage. Many organizations need to understand better the different branches within AI, the varied use cases, and the implementation approaches. It is akin to introducing sport to your child — there are broad benefits to adding a sport to a child's life, but you must choose a specific one to determine the approach, buy the right equipment, and achieve your desired outcome.

This brief article aims to demystify AI by breaking it into key sub-divisions and exploring the various use cases and data dependencies that make these technologies so powerful.

Artificial Intelligence is an Umbrella Term

Artificial Intelligence (AI) is a broad field of computer science focused on creating systems capable of performing tasks that typically require human intelligence. There are three main divisions:

• Machine Learning (ML): A subset of AI that enables systems to learn and improve from experience without being explicitly programmed. Use cases include predictive analytics, optimization and recommendation systems, and fraud detection. Creating the right model using the right approach, training it, and monitoring its drift over time is a complex undertaking led by Data Scientists.
• Deep Learning (DL): A subset of ML that uses neural networks with many layers to analyze various types of data. Use cases include image and speech recognition, autonomous vehicles, and natural language processing.
• Generative AI (GenAI): Algorithms that can generate new content, such as text, images, programming code, and music, based on the data they have been trained on. Large Language Models (LLMs) are a form of GenAI focussed on text generation and provide opportunities for organizations to create their own LLM based on their content.

While most organizations will use DL and GenAI tools and may create their own LLMs, they should be pursuing ML. However, ML depends on data.

The Importance of High-Quality Data

AI systems are only as good as the data they are trained on. High-quality, well-labelled data is crucial for accurate and effective AI models.

• Supervised Learning: Involves training an AI model on a labelled dataset where the desired output is known. Examples include spam detection and image classification.
• Unsupervised Learning: Involves training with unlabeled data to find hidden patterns. Examples include customer segmentation and anomaly detection.

Unsupervised learning identifies patterns and correlations in data, whereas supervised learning seeks a targeted outcome based on a set of inputs. You have the inputs and a desired output; you just don't know how the two are connected.

Effective AI implementations depend on robust data architecture, including data collection, storage, processing, and management systems. Data quality and accessibility are critical prerequisites, especially for real-time or near-real-time solutions. For example, suppose you have two systems but store a person in one system as name and address and the other as first name, last name, address 1, address 2, city, province, and postcode. In that case, this is an ML project by itself, bringing the data together, often with a degree of imperfection. Joining multiple systems with numerous and compounding degrees of imperfection results in something that may be wrong more than it's right.

Conclusion

AI is a multifaceted field with numerous sub-divisions beyond generative AI. GenAI is a powerful time-saving tool, much like a calculator or Excel, but requires training and examples. Staff will not have the innate skills to use GenAI productively; it will be used as a Google alternative.

Other AI areas offer unique opportunities for business value creation driven by high-quality data. To harness AI's full potential, businesses should explore beyond generative AI and invest in robust data architectures to support their real AI initiatives.

GenAI is table stakes. ML is where the real value lies, but you need to feed it timely, high-quality data.