Intro to AI Agents: Metaverse for Education Newsletter Issue #24

Get a generative AI agent to do your boring work for you.

This monthly newsletter serves as your bridge from the real world to the advancements of web3 & the metaverse, specifically contextualized for education. If you haven’t already subscribed, join here:

---

📅 Ed3 Events

• Metaverse Book Club Kick Off (May 6, 2023, Zoom & Ed3verse)

• Ed3 DAO Town Hall (May 20, 2023, Ed3verse)

• The first Edcamp in the Metaverse (May 20, 2023, Ed3verse)

• Web3 & Education writing group (May 28, 2023, monthly, Zoom)

• Thinking with AI: Optimize AI for you (May 30, 2023, WorkAdventure)

• Weekly coffee chat on web3 in education (every Saturday, Discord)

Dear Educators & Friends,

For the last few weeks, I was away from the Edtech world planning and implementing a short term project with intangible profit margins called, ‘my wedding’. The product was successfully shipped in both English & Hindi and has proven to be a customer favorite. As you can see in the video above of my henna, you all were with me in spirit 😊.

As I return from my hiatus, I am impressed by the uptake of AI models and tools in education. Sal Khan’s experiments are the best example I’ve seen of the potential of AI in our current education system. I’m also seeing teachers increasing productivity and decreasing administrative tasks, allowing them to spend more of their time with students, which currently averages 49%.

One of the more exciting developments in this regard is the emergence of autonomous AI agents 😎. Although fundamentally old technology, AI agents have leveraged GPT-4 to simulate human intelligence, behavior, and communication.

Imagine asking an autonomous AI agent to grade an assignment for you. With only the assignment in-hand, it will create a set of action steps to complete this task, create a rubric for grading, determine all acceptable answers, create a repository of feedback, and then…grade the assignment.

This new type of agent is different from what’s existed:

1. Simple reflex agents react to a given situation based on a pre-defined set of rules. A thermostat is a simple reflex agent that triggers heating or cooling based on your pre-defined temperature.

2. Model-based reflex agents assess the current situation and use a model of the environment to make decisions. A self-driving car makes decisions based traffic signals, rules, and other cars.

3. Goal-based agents are programmed with specific targets to achieve and take actions to achieve them. A virtual assistant like Alexa finds information based on user input.

4. Utility-based agents consider the potential outcomes of their actions and choose the action that will result in the best outcome. A financial investment agent considers potential outcomes of different investment options based on pre-defined criteria.

5. Learning agents adapt to new situations and learn from their experiences to improve their decision-making abilities. A chatbot like ChatGPT uses machine learning to improve it’s responses.

Here’s what’s new about the autonomous AI agent:

1. Task-driven autonomous agents use various sets of data and generative AI models to decide by themselves how to attain certain goals, and then execute those tasks.

AI agents can analyze large amounts of data, recognize speech, understand natural language, and perhaps even generate creative content. They will ask themselves all questions required to complete a task and execute it for you.

So where does this leave the teaching profession? More human, actually. We no longer need educators or parents to be administrative assistants, operations managers, attendance takers, and paper graders. Educators can focus on social, intellectual, and emotional relationships with students through hands-on experiences that extend beyond transactional interactions.

Let’s dive into how AI is continuing to change education, and how AI agents can help.

Warmly yours,

Vriti Saraf & k20 Educators

---

Learn about AI Agents

• What are AI agents?

• What is generative AI?

• Researchers populated a tiny town with generative AI agents & it was very wholesome

• Generative AI for the masses

• 📹 The story of babyAGI, the world’s first task-driven autonomous agent

• The anatomy of AI autonomy

• Difference between AI and agents

• AI agents will speed up generative AI race

• The potential of generative AI agents in education

• A very advanced 2002 paper on intelligent agents in education

• 🔉 Experimenting with AI agents

• Will AI help or replace teachers?

• Will universities return to oral exams as AI becomes more popular?

• A fun diversion: AI tasked to destroy humanity

Do Something with AI Agents

• A user friendly AI agent

• Another no-code AI agent

• A personal AI agent

• If you know code, get started with these three:

  • The original autonomous IA agent

  • An iteration on the original

  • A powerful version

📅 Ed3 Events

• Metaverse Book Club Kick Off (Sat. May 6, Zoom & Ed3verse)

• Ed3 DAO Town Hall (Sat. May 20, Ed3verse)

• The first Edcamp in the Metaverse open to all educators (Sat. May 20, Ed3verse)

• Web3 & Education Writing Group (Sun. May 28, monthly, Zoom)

• Thinking with AI: Learn how to optimize AI (Tue. May 30, WorkAdventure)

• Weekly coffee chat on web3 in education (every Saturday, Discord)

---

Mint an Ed3 Educators NFT for life-time access to web3 events, content, conferences, & other perks… and to support the work of this newsletter.

This week’s Metaverse for Education newsletter is about AI agents. More to come on other web3 & emerging tech topics. We’re excited to bring education into the Metaverse & help you leverage the opportunities in the new world.