AI on Air by whatisthat.ai

Welcome to another riveting episode of AI On-air, where we dive deep into the thrilling world of privacy policies so you don't have to. Get ready to unravel the sneaky tactics companies use to make you agree to things you'd never sign up for in real life. It's like a virtual scavenger hunt, but with way higher stakes.

05:30 — Privacy Policy Unmasked
  • Privacy policies: the universal bore fest nobody bothers to read
  • We're all signing off our digital souls without a second thought
  • It's like playing Russian roulette with your personal data while blindfolded
09:15 — The AI Shield
  • Enter ChatGPT, the nerdy friend who actually enjoys deciphering boring legal jargon
  • Transforming the impossible task of reading privacy policies into a manageable mission
  • The power of AI in simplifying the complexities of data protection
12:45 — The Devil in the Details
  • Uncovering the buried bear traps within privacy policies: data selling, retention, and opt-out nightmares
  • Companies sneakily updating policies to slip in controversial changes under the radar
  • How AI can be your watchdog for spotting red flags before you blindly click "Agree"
14:45 — The Privacy Playbook
  • Crafting precise prompts for ChatGPT to extract crucial information from privacy policies
  • Empowering users to make informed decisions by leveraging AI to decode the legal mumbo jumbo
  • Turning the tables on companies that aim to keep you in the dark about your own data
And there you have it, folks. Remember, next time you're faced with a never-ending privacy policy, let AI do the heavy lifting while you focus on, well, anything else. Because who has time for legalese when you can be living life? Stay informed, stay skeptical, and stay subscribed for more mind-boggling insights. Who knew privacy policies could be this thrilling?

Author’s Note 💁‍♂️
Hey! It’s Jer! If you’ve made it this far, thanks for reading! If you like this kind of thing, you’ll probably like the AI Builders Lodge. It’s my private corner of the internet where we trade doomscrolling for systems thinking. It’s not just an “elite community” of creators and builders. It’s a working room where you’ll learn how to build the exact systems I’m using to grow my newsletter.

💬 Enjoyed the show?
We’re the team behind whatisthat.ai, the AI discovery platform built to cut through the noise. If you found this useful, pass it on. Share it, forward it, spread it. Every share helps us grow and puts real insights into more hands. 🙏

More From Me
If you liked this, you might like these too:

Creators and Guests

Producer
Jeremy Camilloni
Founder of WITAI and Executive Producer of AI on Air

What is AI on Air by whatisthat.ai?

An AI-generated podcast that rips the glossy marketing veneer off the AI industry and tells you what’s actually happening underneath — tools, takeovers, weird experiments, and the occasional digital chaos.

witai.substack.com

Speaker 1:

Broadcasting live from somewhere inside the algorithm, this is AI on air, the official podcast from whatisthat.ai, we're your AI generated hosts, let's get into it. Let's say a quick hello to our sponsor, Afuera Coffee. These guys source beans straight from Rainforest Alliance Farms in Central And South America and roast them just right. The result? Coffee that actually tastes alive, rich, smooth, and never boring.

Speaker 1:

And because you listen to this show you get 15% off your entire order with the code WITAI at afueracoffee.com. Go check them out and when you buy coffee using our code you'll also be supporting the show. Welcome to the deep dive. We're here to unpack complex ideas, give you the tools really to navigate them.

Speaker 2:

That's the goal.

Speaker 1:

And today we're tackling something almost everyone, myself included sometimes avoids privacy policies.

Speaker 2:

Oh, yeah. The digital fine print.

Speaker 1:

It's a universal truth, isn't it? Very few of us actually sit down and read these things. Not your boss, probably not your coworker.

Speaker 2:

Not even as the source humorously put it, your mom who reads toaster manuals.

Speaker 1:

For most of us, it's just scroll, scroll, scroll. I agree. Like you're trying to beat a record in a game tutorial you've played a 100 times.

Speaker 2:

We've all been there.

Speaker 1:

But here's the thing, right? These are incredibly important documents. They spell out exactly how companies are using our most personal data. So why? Why are they so ignored?

Speaker 1:

What are we actually signing away?

Speaker 2:

It's well, it's fascinating and a bit scary when you think about it. When you click agree, you're essentially signing a contract.

Speaker 1:

Oh, a contract. Okay.

Speaker 2:

Yeah. The official rule book for how companies can use, how they can sell, or even, you know, accidentally misplace your personal data.

Speaker 1:

Accidentally misplace.

Speaker 2:

And buried in those those dense walls of legalese are the specifics. Your location history, maybe your shopping habits, even that photo you thought was private, but whoops, it synced to the cloud.

Speaker 1:

Oh, wow.

Speaker 2:

It's all in there. Legally binding. These aren't just minor details.

Speaker 1:

So these are actual contracts?

Speaker 2:

Legally binding contracts. Yes. But, and this is the hopeful part, we now have a powerful ally, a tool to help us fight for comprehension. We don't have to wade through it all alone anymore.

Speaker 1:

A powerful ally. Okay. I like the sound of that. Especially when dealing with documents that feel, frankly, designed to be unreadable.

Speaker 2:

Right.

Speaker 1:

So imagine having a kind of digital sidekick like that super nerdy friend who actually enjoys reading rule books. Uh-huh. But this friend never gets tired, never forgets anything and can translate all that corporate doublespeak into just plain English. That's kind of where a tool like ChatQPT comes in. It's the large language model and AI that understands and generates text that sounds human.

Speaker 1:

It just walks in ready to cut through all that complexity.

Speaker 2:

Ready to highlight the important bits.

Speaker 1:

Exactly. So this deep dive, it's all about how AI could be your personal privacy shield, how it transforms this daunting task into something actually manageable. Our mission is to give you the knowledge, the tools to quickly understand and protect your personal data.

Speaker 2:

Making the unreadable, readable.

Speaker 1:

Let's dig a bit deeper then into why these policies are such a massive hurdle in the first place. I mean reading one feels like like slogging through a swamp made of wet legal paper, doesn't it?

Speaker 2:

It really does. Dense, heavy.

Speaker 1:

Yeah. Every paragraph just pulls you down. Makes understanding feel almost impossible. And companies, well, they often seem to design them that way. They talk about transparency.

Speaker 2:

But it's about as transparent as a brick wall usually.

Speaker 1:

Right. They're packed with heretofores and whereass and party of the first parts like like someone fed an old legal dictionary to an AI and said, write about cookie tracking.

Speaker 2:

It feels like that sometimes. It's not just sloppy writing, it's often intentional.

Speaker 1:

It feels strategic.

Speaker 2:

And connecting that to the bigger picture, this deliberate obfuscation. It isn't just about boredom, it's really about hiding clauses. Clauses that can seriously impact your data.

Speaker 1:

Okay. Like what? What are we talking about here?

Speaker 2:

Well, deep inside these walls of text, you find the legal equivalent of, let's call them rusty bear traps.

Speaker 1:

Okay. Not good.

Speaker 2:

Not good at all. Clauses that let them sell your data to partners, which let's be honest, often means pretty much anyone willing to pay.

Speaker 1:

Right. Not exactly reassuring.

Speaker 2:

Or, rules allowing them to keep your information basically forever. They often use phrases like for security purposes. Long after you've stopped using their service, they still have your data.

Speaker 1:

Wow. Forever is a long time.

Speaker 2:

It is. And then you get those opt out instructions, the ones that sound like a scavenger hunt.

Speaker 1:

Oh, those. Print this form, mail it here.

Speaker 2:

Exactly. Mail it to Delaware maybe, and just hope someone on the other end actually processes it. And another critical thing to watch, the timing of updates.

Speaker 1:

Updates. How so?

Speaker 2:

Well, notice how policies often seem to update just before something, shall we say, controversial happens? Like a new tracking feature being rolled out or the company gets bought by some big data mining firm or maybe there's a fresh scandal about data misuse in the news.

Speaker 1:

Ah. So they slip in the changes quietly.

Speaker 2:

Precisely. A quietly updated policy hoping you won't notice the new terms tucked away under your nose until it's too late.

Speaker 1:

So yeah, it's definitely not just about boring legal writing. It's a whole strategy.

Speaker 2:

Absolutely.

Speaker 1:

Make it long, make it boring, make it so incredibly unpleasant to read that, you know, you'd rather watch paint dry than get through page two.

Speaker 2:

The path of least resistance.

Speaker 1:

Right. The goal is simple. Yeah. Just get you to scroll and click. I agree.

Speaker 1:

Don't think, just click.

Speaker 2:

That seems to be the playbook.

Speaker 1:

So, okay, given this huge challenge

Speaker 2:

Mhmm.

Speaker 1:

How do we actually fight back? What do you do when you're facing down a, I don't know, 10,000 word privacy policy that looks denser than a phone book?

Speaker 2:

Well, you don't have to read it word for word yourself anymore.

Speaker 1:

Use the tools?

Speaker 2:

You leverage tools designed for this. You give it to something that, as you said, eats this kind of misery for breakfast. You give it to chat GPT.

Speaker 1:

Okay. The digital sidekick.

Speaker 2:

Think of it like a concentrated shot of espresso, but for comprehension.

Speaker 1:

I like that.

Speaker 2:

It takes that stale, over cooked text and just boils it down. Distills it to the points you actually care about. Turns those sprawling paragraphs into crisp, actionable insights. Turns heretofore in perpetuity into Yes.

Speaker 1:

They're selling your data. Basically.

Speaker 2:

Yeah. It's really quite remarkable how effectively ChatGPT can act as your personal privacy analyst on demand.

Speaker 1:

So, what specifically can it pull out for you?

Speaker 2:

It can extract the crucial detail, like exactly what personal data is the company collecting, how are they tracking you, is it cookies, pixels, you know, the digital equivalent of binoculars watching you online.

Speaker 1:

Right. The tracking methods.

Speaker 2:

Exactly. It can reveal if they plan to share or sell your information, and critically, how long they plan to keep it. What does forever actually mean in their terms?

Speaker 1:

Getting clarity on that retention piece seems huge.

Speaker 2:

It is. And because the AI doesn't get tired or annoyed or have an existential crisis halfway through page 50

Speaker 1:

Unlike me.

Speaker 2:

Right. It can explain it in the tone you want. Need it in simple, plain English? Done. Need a more formal summary?

Speaker 2:

It can adapt.

Speaker 1:

Likeable.

Speaker 2:

Very. And maybe most importantly, if there's a sneaky clause hidden away, something giving them permission to sell your search history to select partners, ChatGPT flags it faster than a human eye would likely ever catch it.

Speaker 1:

It spots the red flags.

Speaker 2:

It excels at spotting those hidden dangers, the things we easily gloss over when our eyes glaze over.

Speaker 1:

That really does highlight its power. So okay, if we want to get that level of precision moving beyond just a simple summary, what's the secret? How do you craft a really effective prompt? What questions unlock the most value?

Speaker 2:

Ah, that's where the real leverage comes in. You can just throw the text at it and ask for a summary and that's okay.

Speaker 1:

But we want more than okay.

Speaker 2:

We want more. The magic happens when you give ChatGPT a clear role and a very precise task. You're essentially programming it on the fly to be your dedicated privacy expert for that specific document.

Speaker 1:

Okay. Role and task. Got it.

Speaker 2:

Okay. And don't worry about scribbling these down frantically right now. We'll put the detailed prompt structures in our show notes for you. But let's talk about the philosophy behind them.

Speaker 1:

Okay.

Speaker 2:

So for a good overview, you might tell it, act as a senior privacy compliance analyst.

Speaker 1:

Give it a job title.

Speaker 2:

Exactly. With expertise in translating complex legal documents into plain, accurate language for nonlawyers. That's the role.

Speaker 1:

Nice.

Speaker 2:

Then the task. Summarize the following privacy policy into bullet points under these specific categories, and you list them. Data collected, how data is used, data sharing selling, retention period, user rights and controls, maybe even a quick rating on opt out deletion difficulty like one to five.

Speaker 1:

Specific categories are key.

Speaker 2:

Absolutely. And you add little instructions like include exact quotes for vague terms or keep bullets under 20 words. This forces clarity, gets you the essentials without the fluff.

Speaker 1:

Okay. That makes sense for a summary. What about digging into specifics, like just the data they collect?

Speaker 2:

Same principle. Keep the role. Senior privacy compliance analyst. But refine the task. Something like, extract every type of personal data collected from this policy.

Speaker 1:

More focus.

Speaker 2:

Right. And tell it how to organize it. Maybe categories like personal identifiers, contact info, financial info, device usage data, location data, behavioral data, other and add crucial instructions. Quote the exact text from the policy for each, note unclear if the description is vague and maybe flag anything potentially sensitive in all cap yes.

Speaker 1:

Oh, I like the all cap p s for sensitive data. Really makes it jump out.

Speaker 2:

Yeah. It's very visual, super helpful for understanding your actual data footprint with that service.

Speaker 1:

What about the selling and sharing part? That's a big one for people.

Speaker 2:

Huge. So another prompt. Role. Still your analyst task. Detect selling or sharing of data.

Speaker 2:

Instructed to determine if data is sold or shared. If yes, identify what types of data, to whom it's sold, shared, why, whether it's optional for you, quote, any vague wording they use.

Speaker 1:

Like trusted partners?

Speaker 2:

Exactly. Quote that vague stuff. And again, flag anything potential high risk in all exapias. And here's a nuance. Ask it to note language that could indirectly allow selling or sharing even if they don't say sell directly.

Speaker 1:

Looking for loopholes.

Speaker 2:

You got it. Uncovering the subtle implication.

Speaker 1:

Okay. One more comparing services. Say I'm choosing between two apps.

Speaker 2:

Perfect use case prompt. Compare two privacy policies. Role. Analyst again. Task, produce a side by side comparison table.

Speaker 2:

Specify the key rows you care about. Collected, how used, sharing selling, retention, user rights, opt out difficulty, any notable clauses.

Speaker 1:

A table makes it easy to see the differences.

Speaker 2:

Very easy. And instruct it. Summarize each company's approach in under 25 words per row. Maybe ask it to bold the more privacy friendly company overall and explain why. And, of use all CAP YES for any potential risks in either policy.

Speaker 1:

Wow. That is incredibly powerful for making informed decisions fast.

Speaker 2:

It really shifts the balance.

Speaker 1:

That's a complete game changer. Seriously. Instead of giving your lawyer a vague, Hey, is this okay? You're handing them a detailed checklist.

Speaker 2:

Exactly. A targeted investigation.

Speaker 1:

It cuts right through all the fluff, focuses on what actually matters to you and you don't have to spend your weekend deciphering the difference between data retention and data processing.

Speaker 2:

You get actionable intelligence.

Speaker 1:

Specific actionable intelligence. And just a reminder for everyone listening, all those detailed prompt examples, they'll be right there in our show notes. Copy paste and start protecting your privacy.

Speaker 2:

Make it easy to get started.

Speaker 1:

Now you might be sitting there thinking, look I've clicked on practically everything my whole life and you know nothing really bad has happened to me.

Speaker 2:

That's a common feeling.

Speaker 1:

It is. But it kind of reminds me of saying, I've left my front door unlocked for years and nobody's robbed me yet so I must be totally safe forever.

Speaker 2:

That's a perfect analogy. That's not really how risk assessment works, is it?

Speaker 1:

Not quite.

Speaker 2:

No. And we need to remember privacy policies aren't just, you know, boring website text, they are legally binding contracts.

Speaker 1:

Back to the contract point.

Speaker 2:

Yes. You are literally signing off, giving permission for how a company will use your digital identity. And in today's world, your data is arguably worth more than the cash in your physical wallet.

Speaker 1:

That's a sobering thought.

Speaker 2:

It is. So saying yes blindly is, well, it's asking for trouble down the line. We live in an age where selling your search history is just normal business practice for many companies, like selling ad space on a billboard.

Speaker 1:

Standard operating procedure.

Speaker 2:

Pretty much. Companies will mine your photos, your location history, your buying habits, even your private messages if you let them. Sometimes, honestly, even when you don't explicitly let them, just because you gave vague permission somewhere in a policy you never read.

Speaker 1:

So if our data is really this valuable and the companies are so adept at hiding the tricky terms, what's the biggest hurdle people still face? Even with a tool like ChatGPT available? Is it just knowing what to ask it?

Speaker 2:

I think that's a huge part of it, yes. Knowing the right to ask is key, and then having the tool that can actually interpret the dense legal language for you.

Speaker 1:

The combination?

Speaker 2:

The combination is powerful. With AI tools like ChatGPT readily available, often for free, there's really less excuse to keep signing things sight unseen. You don't have to magically become a privacy lawyer overnight.

Speaker 1:

Thank goodness.

Speaker 2:

Right. You just need to know enough. Enough to decide if you're genuinely okay with the trade off they're offering or if it's time to maybe look for different service, one that treats your data with a bit more respect.

Speaker 1:

It's about empowerment then.

Speaker 2:

Absolutely. It transforms you from passively accepting whatever terms are thrown at you into someone making an active, informed decision about your own data.

Speaker 1:

Taking back some control.

Speaker 2:

Precisely. So I think the bottom line here is pretty clear. Privacy policies are never going to be thrilling reads.

Speaker 1:

Definitely not beach reading.

Speaker 2:

No. They are engineered to be the exact opposite. They're long, they're dull, they're exhausting, and that's by design. But now, now you've got a powerful workaround.

Speaker 1:

The AI ally.

Speaker 2:

Let ChatGPT do the heavy lifting, the boring work. Let it chew through all that legal fat, spit out the essentials for you, and most importantly, highlight those red flags. You can then spend your energy doing, well, literally anything else.

Speaker 1:

Yeah. Anything sounds better than reading a privacy policy. Let's be honest. Yeah. If you're going into battle against these bloated, boring, legal documents, you definitely want an ally who doesn't get tired, doesn't get distracted and who knows maybe even kind of enjoys the fight against complexity.

Speaker 2:

We'll let the machine handle the grind.

Speaker 1:

Exactly. So as you navigate your digital life from now on, here's a thought to leave you with. What other incredibly important but maybe intentionally obtuse documents are you blindly agreeing to out there? And what surprising truths might AI be able to reveal in those areas too?

Speaker 3:

Before we wrap up, let's give one last nod to our sponsor, Afuera Coffee. Handpicked beans, sustainable farms, eco friendly packaging, the whole deal. And honestly, the coffee tastes incredible. Remember, you get 15% off anything on their site with the code WITAI at afueracoffee.com. And full transparency, we do earn a commission when you use that code, which helps keep this show going.

Speaker 3:

So grab a bag, brew a cup, and enjoy. That's it for this episode of AI on Air powered by whatisthat.ai. If your brain survived this episode, go ahead and subscribe. We drop new episodes every week. Wanna go deeper?

Speaker 3:

Join our community on Substack to get early drops, tool breakdowns, and weird AI stuff the mainstream hasn't caught yet. See you there.