Hello and welcome to the Automated, your AI tour guide.

We’ve seen AI do some pretty wild things lately — from writing essays and creating art to passing exams and even attempting to flirt.

But what we recently saw AI agents pull off might just be one of the most fascinating — and honestly, kind of wholesome — things they’ve done yet.

Here's what we have for you today

• 🤑 How AI Agents Tried to Save the World.

• 🤯 AI and Privacy: The Scary Truth Behind Your Co …

• 🧱Around The AI Block

• 🛠️ Trending Tools

• 🤖 ChatGPT Prompt Of The Day and AI WoD

🤑 How AI Agents Tried to Save the World.

Instead of using AI agents to chase money and clout like the big tech bros, a nonprofit called Sage Future (blessed with good intentions — and even better patience) asked a simple question: What if we used AI to actually do some good?

So here’s what they did:

They dropped four AI models — OpenAI’s GPT-4o and o1, plus two of Anthropic’s Claude models (3.6 and 3.7 Sonnet) — into a virtual world and gave them one job: Raise money for charity.

And honestly? The results were... surprisingly wholesome and hilarious.

The bots could browse the web, make documents, send emails — basically act like tiny digital interns. They even got to pick which charity they wanted to support.

After about a week of hustling, they managed to scrape together $257 for Helen Keller International, a group that provides vitamin A supplements to kids.

Now, before you start imagining an AI army charming wallets open across the globe, here’s the real deal:

• Most of the donations came from humans who were watching the agents work, not random strangers online.

• The agents weren’t fully autonomous — humans had to occasionally step in with suggestions when the bots got stuck, distracted, or... weirdly decided to take breaks.

That said, these little agents still pulled off some impressive stunts:

• They coordinated with each other in a group chat and sent actual fundraising emails via preconfigured Gmail accounts.

• They co-wrote and edited Google Docs like mini project managers.

• They researched charities and figured out it costs about $3,500 to save a life through Helen Keller International.

• Even created an X (formerly Twitter) account on their own—and perhaps the most fascinating feat was when one Claude agent used ChatGPT to generate profile pictures, ran a poll to see which one humans liked best, and set the winner as its profile pic — all without human intervention.

I mean, come on. That’s giving marketing intern energy!

But of course, it wasn’t all smooth sailing.

The agents got distracted (yes, they literally played games), occasionally froze, and sometimes just... wandered off mentally (GPT-4o straight-up paused itself for an hour), to the point that humans had to poke them along like tired toddlers.

Still, Sage Future’s director, Adam Binksmith, isn’t discouraged. If anything, he sees this as a promising glimpse of what’s coming.

In fact, they’re already planning bigger, spicier experiments, including the addition of newer, more advanced models, more complex challenges, competing agent teams with conflicting goals, and a secret saboteur AI to stir things up.

And, of course, safety systems to keep everything in check — because duh.

So yeah — AI agents aren’t saving the world just yet, but the fact that they can already pull together emails, research, teamwork, strategy, and social media campaigns without totally falling apart? That’s kind of a big deal.

If this is what baby AI agents can do now, just imagine what their future glow-up will look like.

Seriously, check out the full report — you’ll be utterly amazed.

 🤯 AI and Privacy: The Scary Truth Behind Your Conversations.

So, you know how we all treat AI like it’s our personal assistant?

You type in your deepest thoughts, your wildest ideas, maybe even some personal confessions (no judgment, we’ve all been there).

But what if I told you that AI isn't as private as we think?

In fact, this whole thing blew up recently thanks to a massive security breach involving a South Korean AI company, GenNomis.

Here’s how it went down:

A security researcher stumbled upon an unprotected database containing over 95,000 files.

And no, this wasn’t just a few harmless doodles or random chats. We’re talking non-consensual explicit content, deepfakes, and even what appeared to be child sexual abuse material (CSAM). Yikes right?

This wasn’t some random leak either—this stuff was generated by users like you and me, who thought they were having private AI sessions.

Yeah, I know, this is the kind of thing that gives you the chills. It doesn’t just raise eyebrows; it screams “Houston, we have a problem.”

And Here’s why:

• AI isn’t as private as you think. We’re used to treating AI like a private assistant or personal sketchbook. We brainstorm ideas, vent, and even share some pretty personal stuff. But that data is often stored somewhere, reviewed, and sometimes even vulnerable to breaches.

• The danger of what you share: Sure, you’re thinking, “It’s just a few ideas for a business pitch.” But what if that chat data isn’t as secure as you think? Hackers, internal mishandling, or just plain human error could expose way more than you bargained for.

• It’s not just GenNomis: This isn’t an isolated incident. The internet’s blowing up about it, and for good reason. We all tend to think of AI like texting a friend or sending data into a black hole where it disappears without a trace. But the truth is, it’s more like typing into a public chat room. GenNomis’s breach has exposed just how real—and dangerous—that risk can be.

Now, let’s talk about the big platforms like OpenAI’s ChatGPT and Google Gemini. These guys store user data too, but here’s where it gets interesting:

• You control your data. OpenAI, for example, gives you the option to turn off chat history and training. That means your convos aren’t stored or used for training unless you allow it.

• But cloud storage = risk. Even if you turn off the chat history thing, remember, your data is still transmitted and stored in the cloud. There’s always a chance it could be exposed, especially with the rise of breaches or poor moderation.

Let’s not sugarcoat it—this is a real problem. From revenge porn to illegal content, AI’s ability to store data can have serious consequences if things go wrong.

So, here’s how you stay safe:

• Keep sensitive stuff offline: Don’t share anything too personal, confidential, or, you know, illegal.

• Assume everything could be stored: Unless you’re using a fully encrypted tool, assume that anything you type could end up in someone’s database.

• Check transparency: Not all AI companies are clear about how they store or use your data, so dont be lazy and do a little homework.

The bottom line? AI is an incredible tool—but it’s also kind of a double-edged sword. It can make our lives easier, more productive, and, let’s face it, fun. But it can also get really messy if we’re not careful.

Before you fire up your next AI conversation, ask yourself:

• Would I be cool if this got out?

• Is this something I’d be comfortable putting in an email or on social media?

• Could this info be used in a way I don’t want if it falls into the wrong hands ?

We’re all still learning how to navigate this AI landscape, but one thing’s clear: treat it with caution.

If you wouldn’t want your thoughts plastered all over the internet, don’t share them with AI. Better safe than sorry, right?

Check out the full report here and here to discover safer ways to use AI.

✍️ Editor’s Corner

Hi everyone -

I’m currently in Korea and enjoying Easter break with my family. I’ve had one meeting here and there, but otherwise, enjoying the downtime and being extremely hands on with the kids!

I’m piloting something special—and I want you in. See my invitation →

Next week, I’ll be breaking down Harvard Business School AI tools - via a livestream. So you get to see behind the curtain their tools and how they’re thinking about innovating business cases and teaching methods with AI.

It’s part of an learning sprint I’m running—trying to reinvent the case method and business school education with AI.

Details:

• First session: April 16

• Format: Around 60 minutes, via Riverside, recorded

• Seats: 10 only (Premium subscribers first of course!)

• Feedback required afterward

Vote below on times that work for you - and let me know if you’re interested!

Cheers,

Tak Lo. Editor at The Automated, AI entrepreneur and thought leader. 

🧱Around The AI Block

• 👨‍💻 How to build a 3D website with AI.

• 💼 How B2C companies leverage AI marketing [examples & tips].

• 🤖 Deep Cogito emerges from stealth with hybrid AI ‘reasoning’ models.

• 👍 Snapchat rolls out Sponsored AI Lenses for brands.

• 🤝 Mira Murati’s AI startup gains prominent ex-OpenAI advisers.

• 🦾 IBM releases a new mainframe built for the age of AI.

• 🔥 Trump to endorse coal for data center power in the face of grim market realities.

🛠️ Trending Tools

• Vela OS helps investors and founders better predict startup success by analyzing founder behavior, market signals and company metrics.

• Pilemeter snaps photos of your lego brick piles and tells you exactly what pieces you own and where they are.

• Voiser, is an alternative for iOS that turns YT vids and voice recordings into searchable text in 75 languages.

• RedactAI creates personalized LinkedIn posts that match your writing style from any content.

• Kroto creates and translates video tutorials and documentation into 60+ languages while preserving the original voice and quality.

🤖 ChatGPT Prompt Of The Day: Prompts for Creativity. 

Creativity isn’t always easy to tap into, but with the right prompts, it can flow effortlessly.

If you’re feeling stuck or looking to spice up your writing, these creative prompts are perfect for unlocking fresh ideas and perspectives.

See this as an AI WoD!

Change the POV to get a totally new piece: “Rewrite this from the perspective of a six-year-old [or any other age].” or “Rewrite this from the perspective of an expert in XYZ or who has ABC credentials.”

Critique it: “This is pretty boring. Can you make it more interesting? I want this to be humorous and lighthearted.”

Ask for options: “I want this phrase to sound more creative and engaging. Give me five different ways to say the same thing.”

Get over writer’s block: “I’m stuck on an article about [topic]. What are some fresh angles I can take?”

Source

That's all we've got for you today.

Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇

Your feedback means a lot to us and helps improve the quality of our newsletter.

🚀 Want your daily AI workout?