AI Primer

GenAI assists in content creation & marketing by improving the scale and efficiency of content production. GenAI models create new content, including text and visuals, from user-defined inputs and established brand parameters. This capability allows marketing teams to generate multiple content iterations—such as ad copy, email drafts, and preliminary outlines—more quickly than traditional methods.

Furthermore, GenAI can process customer data to customize content for specific audience segments, facilitating more targeted communication. This shifts production toward the creation of individualized customer experiences and contributes to higher overall marketing efficiency.

GenAI is enhancing Customer Service by enabling sophisticated, conversational interactions. GenAI models allow virtual agents to understand and respond to the nuance and complexity of natural human language. This technology processes unstructured customer input from sources like email, chat, and voice transcripts to accurately interpret intent and sentiment. Consequently, GenAI can generate personalized, context-aware resolutions, summarizing issues for human agents, or autonomously handling a higher percentage of complex inquiries without escalation, leading to improved service consistency and efficiency.

Bridging the cognitive gap left by traditional, rule-based Robotic Process Automation (RPA). While RPA excels at automating structured, repetitive tasks, GenAI brings intelligent capabilities by handling unstructured data (like emails, documents, or voice notes) and generating original content or decisions.

This integration allows for end-to-end intelligent automation, where GenAI can summarize customer inquiries, extract key data from a contract, generate a personalized response, and then trigger the RPA bot to execute the next transactional step, leading to fully unattended process automation in complex workflows such as customer service, accounts payable, or compliance reporting.

GenAI can assist in rapidly exploring vast solution spaces. Instead of relying solely on iterative human adjustments, GenAI tools are used to generate novel and optimized design options that meet predefined constraints and performance targets. In physical product design, this involves Generative Design, where the AI creates optimized geometries for lighter, stronger parts (e.g., in aerospace).

GenAI can analyze massive datasets to predict molecular properties or simulate complex interactions (e.g., drug discovery), generating new candidate molecules for synthesis or testing. This capability speeds up the initial concept phase, reduces physical prototyping cycles, and uncovers non-intuitive solutions, significantly shortening the time from concept to viable product or discovery.

GenAI can augment the capabilities of engineers and accelerate various stages of the development lifecycle. This involves automating the generation, completion, and modification of code, documentation, and tests. GenAI acts as an intelligent pair programmer by analyzing existing codebase patterns, natural language instructions, and desired outputs to suggest or write functional code snippets, translate code between languages, and identify potential bugs. This allows developers to focus on higher-level system architecture and complex problem-solving rather than boilerplate or repetitive coding tasks, leading to faster development cycles and improved code consistency.

Chatbots such as Google Gemini and ChatGPT use generative AI and natural language processing to simulate human-like conversations in a chat window where the user can ask the bot to help with a variety of tasks, including editing or writing emails, essays, code, and more.

The process of customizing a pre-trained Foundation Model using a smaller, task-specific dataset (e.g., your company's proprietary data). Fine tuning improves the model's performance, style, and accuracy for a specific use case.

When an AI model generates an output that is confident but factually incorrect, nonsensical, or made up. It can be a significant limitation, as models prioritize coherence over truthfulness; they are trained to predict the next plausible word.

This integration allows for end-to-end intelligent automation, where GenAI can summarize customer inquiries, extract key data from a contract, generate a personalized response, and then trigger the RPA bot to execute the next transactional step, leading to fully unattended process automation in complex workflows such as customer service, accounts payable, or compliance reporting.

A field of study with a range of approaches to developing algorithms that can be used in AI systems. AI is a more general term. In ML, an algorithm will identify rules and patterns in the data without a human specifying those rules and patterns.

These algorithms build a model for decision making as they go through data. (You will sometimes hear the term machine learning model.) Because they discover their own rules in the data they are given, ML systems can perpetuate biases. Algorithms used in machine learning require massive amounts of data to be trained to make decisions.

Natural Language Processing (NLP) is the field of AI where computer science meets linguistics to allow computers to understand and process human language.

The text instruction or query a user gives to a GenAI model to generate an output. The quality of the prompt often determines the quality of the output (Prompt Engineering).

A framework that allows a model to access, retrieve, and use information from external, trusted knowledge sources (like a company's internal documents) before generating a response. It reduces hallucinations and ensures responses are based on the most current and relevant proprietary data.

A parameter that controls the randomness or creativity of the model's output. A low temperature (e.g., 0.2) makes the output more deterministic and conservative; a high temperature (e.g., 0.8) makes it more varied and unpredictable.

Models read and generate based on sequences of tokens. Costs and context limits are often measured in tokens.

The initial process where a model learns from a vast, diverse dataset to establish its core capabilities (creating a Foundation Model). This is compute-intensive, time-consuming, and expensive.

Addressing the risk of students using AI to bypass critical thinking and original work, which requires updating plagiarism policies and teaching students how to ethically cite and integrate AI tools.

Additional information: Siena University Academic Integrity Policy

AI models can confidently fabricate facts, references, or data (hallucination). All AI-generated content must be rigorously verified against original sources.

The models' outputs can reflect biases present in their training data, which could compromise research objectivity or perpetuate inequities if not carefully monitored.

Researchers must be cautious about uploading proprietary or sensitive data and licensed academic materials, ensuring compliance with institutional policies and publisher agreements.

Recognizing that AI tools are very useful—especially in research—and that its usage will grow, we must be conscious of the environmental impact that comes from its use. NYU is committed to reducing these impacts.