17:48
Generative AI Vs Agentic AI Vs AI Agents
Krish Naik
Overview
This video explains the distinctions between Generative AI, AI Agents, and Agentic AI. Generative AI focuses on creating new content like text or images based on prompts, using large models trained on vast datasets. AI Agents are systems designed to perform specific tasks, often by utilizing LLMs and external tools (like internet searches) through 'tool calls' when they lack direct information. Agentic AI, on the other hand, involves a system of multiple AI Agents collaborating and communicating to achieve a more complex, multi-step goal, potentially incorporating human feedback.
How was this?
Save this permanently with flashcards, quizzes, and AI chat
Chapters
- Generative AI models (like LLMs and large image models) are massive, trained on huge datasets.
- Their primary function is to generate new content, such as text, images, audio, or video, in response to user prompts.
- These models are 'reactive,' meaning they require specific instructions (prompts) to produce output.
- Applications like chatbots that generate new content are examples of generative AI.
Understanding Generative AI is crucial as it forms the foundation for many modern AI applications, enabling content creation and interaction.
Asking a large language model to 'generate a new image related to agentic AI' or to 'act as a data scientist and take an interview.'
- Large Language Models (LLMs) are central to both generative AI and AI agents.
- LLMs are trained on past data and lack real-time information or access to private/current data.
- This limitation means they cannot answer questions about current events or specific, non-public information without external help.
- They can generate new content but struggle with tasks requiring up-to-the-minute or specialized knowledge.
Recognizing the limitations of LLMs highlights the need for additional mechanisms to access and process real-world, current information.
An LLM being unable to provide the result of a recent IPL match or current news because it's not connected to the internet.
- AI Agents are designed to perform specific, defined tasks.
- When an LLM cannot answer a query directly, it can use 'tool calls' to access external resources or APIs.
- These tools can include internet search engines, databases, or other specialized services.
- The LLM then processes the information retrieved from the tool and provides a summarized response to the user.
AI Agents extend the capabilities of LLMs by enabling them to interact with the outside world and retrieve necessary information to complete tasks.
An LLM using a tool like 'Tavily' (an internet search API) to find today's AI news when it cannot access current information itself.
- Agentic AI refers to a system where multiple AI Agents collaborate to achieve a larger, more complex goal.
- This is different from a single AI agent performing one specific task.
- In an agentic system, agents can communicate with each other, passing information and coordinating their actions.
- This allows for the automation of multi-step workflows that would be too complex for a single agent.
Agentic AI represents a significant advancement, enabling the automation of complex processes by breaking them down and having specialized agents work together.
A system that takes a YouTube video URL, extracts the transcript (Agent 1), generates a title (Agent 2), creates a description (Agent 3), and writes a conclusion (Agent 4) to produce a blog post.
- Generative AI creates new content based on prompts.
- An AI Agent performs a single, specific task, often using tools.
- Agentic AI involves multiple AI Agents collaborating to solve a complex, multi-step problem.
- The core distinction lies in the scope of the task and the level of collaboration between AI entities.
Clearly distinguishing these concepts is essential for understanding the current landscape of AI development and for building sophisticated AI applications.
Key takeaways
- Generative AI models are powerful content creators but are reactive and require explicit prompts.
- LLMs, while versatile, have limitations in accessing real-time or private data.
- AI Agents overcome LLM limitations by using 'tool calls' to interact with external data sources and APIs.
- An AI Agent is designed for a specific task, whereas Agentic AI orchestrates multiple agents for complex workflows.
- Agentic AI systems enable collaboration between specialized AI agents to achieve larger goals.
- Understanding the differences helps in choosing the right AI approach for specific problems.
- The ability to connect LLMs to external tools is key to building functional AI agents.
Key terms
Generative AILarge Language Models (LLMs)Large Image ModelsAI AgentsAgentic AIPromptReactive AITool CallAPIWorkflow
Test your understanding
- How does Generative AI differ from AI Agents in terms of their primary function?
- What is a 'tool call' and why is it important for AI Agents?
- Explain the main difference between an AI Agent and an Agentic AI system.
- What are the limitations of LLMs that necessitate the use of AI Agents and external tools?
- How can Agentic AI be used to automate complex, multi-step tasks?