SHSU Generative AI Guidance
Artificial intelligence (AI) is a broad field of computer science focused on creating systems that can perform tasks that would typically require human intelligence. These tasks include learning patterns, problem-solving, decision-making, and understanding data. AI systems work by analyzing vast amounts of data to find patterns and relationships, then using those insights to act or make predictions.
-
Traditional AI
Traditional AI, also known as "narrow" or "rule-based" AI, operates using explicitly programmed rules, logic, and manually created knowledge rather than learning from data. This is an important distinction: Traditional AI systems do not learn patterns from data the way Machine Learning systems do. Instead, they rely on logic, rules, and mathematical reasoning that programmers or domain experts have built into the system.
Examples of Traditional AI in Action:
- Spam Filters: Email programs that follow a set of rules programmed by engineers (keywords, sender reputation, formatting patterns) to identify and block spam emails. The filter doesn't learn from seeing millions of emails; it applies pre-defined detection rules.
- Medical Diagnostic Systems: Traditional diagnostic systems that use expert-created decision rules (if patient shows symptoms A, B, and C, consider disease X). This contrasts with modern AI that learns patterns by analyzing thousands of labeled medical images.
- Credit Scoring: Systems that evaluate loan applications using explicit mathematical formulas and business rules (if income > $X, debt-to-income ratio < Y, then approve). Rules are set by analysts, not learned from data.
-
Generative AI
Generative AI (Gen AI) is a type of AI that can create new, original content such as text, images, audio, video, or code. Unlike traditional AI, which analyzes and classifies existing data, generative AI learns patterns and structures from massive datasets and uses that knowledge to generate entirely novel content in response to a user's prompt or request. It doesn't just recognize patterns it synthesizes and creates. Technically, this is an application layer on top of modern AI algorithms.
Examples of Generative AI in Action:
- Tools like ChatGPT, Google Gemini, and Claude that can write essays, answer complex questions, summarize documents, create code, or engage in nuanced conversations based on text prompts.
- Systems like DALL-E, Midjourney, and Stable Diffusion that create unique, photorealistic, or stylized images from text descriptions.
- AI tools like SunoAI, Loudly, and MusicGen that generate original musical pieces in specified styles, genres, or moods.
- Emerging tools that can create short video clips or animations from text descriptions or existing images like Googles Veo.
- Tools like GitHub Copilot that auto-complete code, suggest functions, or generate entire code snippets based on comments and context.
The distinction between Traditional AI and Generative AI is critical for understanding when each approach is most valuable and how they fit into the broader landscape of AI technologies.
GenAI is NOT a separate technique but rather an APPLICATION that primarily uses:
- Large Language Models (Transformer-based) for text generation
- Diffusion Models for image generation
- GANs for various content generation
- Autoregressive models for sequence generation
Students
Faculty/Instructors
Staff
Run AI locally. Keep your data private. Maintain full control.
This presentation walks through why local LLMs are becoming essential for research, healthcare, and enterprise settings—and how you can deploy them yourself using accessible tools like LM Studio.