Libertatis Learning · Interactive Digital Privacy Route

Content verified on November 12, 2026

Learn digital privacy and protect your personal data online

If you are looking for clear, practical guides to understand how your personal data circulates online and apply concrete changes today, this route brings three pieces together in one place: immediate actions for your accounts and devices, six step-by-step modules with everyday examples, and a verified directory of trustworthy guides.

You do not need prior technical knowledge, an account, or personal data inputs. You can apply one high-impact improvement in five minutes, work through the full course, or practice in the interactive simulator.

Free · No sign-up or tracking of your answers · Session progress saved locally in your browser

Jump straight to the interactive practice lab (4 stations) → →

1. Immediate action · First visit

I want to protect my data right now: 3 high-impact actions in under 15 minutes

You do not need to overhaul every tool at once. Pick just one of these three actions to start today: each card shows where to begin, how to verify it worked, and what it protects (along with its real limit).

Action 013 to 5 minutes

Review a sensitive app permission on your phone

Check which apps have access to your microphone, contacts, or precise location, and restrict any that do not need them for their core function.

Where to start

  1. On Android: open Settings > Security & privacy > Privacy > Permission manager (on some devices, Settings > Apps > Permissions). Open Location, Microphone, or Contacts.
  2. On iPhone (iOS): open Settings > Privacy & Security and tap Location Services, Microphone, or Contacts to see authorized apps.
  3. Pick an everyday app (such as a notes, shopping, flashlight, or game app) and change the permission from 'Allow all the time' to 'Allow only while using the app', 'Ask every time', or 'Don’t allow'.
How to verify the result:: Close the app and reopen it to perform its usual task (such as reading a note or browsing a catalog). Confirm that the app works normally without triggering the microphone or location indicator in your phone’s status bar, or that it asks for confirmation only when you tap a feature that genuinely requires it.
What protection it provides:: Prevents the app from collecting ambient audio, your address book, or your exact GPS coordinates in the background when you are not using a feature that justifies it.
Important limit:: Revoking system permissions does not delete data you already sent earlier, nor does it stop the app from logging what you type or tap inside its own interface.

Official reference documentation

Action 028 to 10 minutes

Protect your primary email account with a strong sign-in method and recovery plan

Your primary email acts as the master key for resetting passwords on banking, social, and cloud accounts. Strengthen how you sign in and make sure you can recover access if you lose your phone.

Where to start

  1. Open the security settings of your primary email provider (for example, in Google Account: Security > How you sign in to Google; in Outlook/Microsoft or Proton: Security section).
  2. Replace any reused password with a long, unique passphrase stored in a password manager, or set up a passkey on your personal device.
  3. Turn on two-step verification (2FA), preferably with an authenticator app or security key, and download or write down your one-time backup recovery codes on paper before signing out.
How to verify the result:: Open a private or incognito browser window, try signing in to your email account, and confirm that the service requires your second factor or passkey. Also verify that your recovery phone number or backup email is still under your control.
What protection it provides:: Prevents a password leaked from another website or a basic phishing attempt from letting an intruder take over your email and associated accounts, while keeping a safe recovery path if you lose or replace your phone.
Important limit:: Protecting the account against intruders does not prevent the email provider itself from processing metadata or content according to its terms of service.

Official reference documentation

Action 033 to 5 minutes

Check who can see a shared cloud link or your social media profile

Many cloud documents and social profiles remain open to 'Anyone with the link' or the general public long after their original purpose has passed.

Where to start

  1. In a cloud document or folder (such as Google Drive, OneDrive, or Dropbox): open the most sensitive file you shared recently, click Share, and inspect 'General access'.
  2. If it says 'Anyone with the link', switch it to 'Restricted' so only authorized accounts can open it, and change collaborators from 'Editor' to 'Viewer' if they no longer need to modify it.
  3. On a social network: open Settings & privacy > Account privacy or Audience and check whether your phone number, email, contact list, or older posts are visible to any visitor.
How to verify the result:: Copy the document link or your public profile URL and open it in a private or incognito browser window (without signing in). Confirm that the document asks for access permission instead of opening directly, and that your profile hides unnecessary personal details.
What protection it provides:: Closes accidental access to strangers, search engines, or third parties who receive a link forwarded outside the original group.
Important limit:: Restricting the audience does not stop an authorized person from downloading, copying, or taking a screenshot of what they are allowed to view.

Official reference documentation

2. Step-by-step learning route · 6 practical modules

I want to learn step by step: understand how your data flows and decide with confidence

You can follow the six modules in order or jump straight to the question you want to solve today. Each module includes an observable objective, a worked example, an interactive decision with choice-specific feedback, and an optional real-world action.

Module 1 of4 min read + interactive exercise

Module 1. What data says about you: provided, observed, and inferred

Why can a platform assign you tastes, income brackets, or routines you never typed into any form?

Observable objective:: Classify information in an everyday app into provided, observed, and inferred data, and explain why an automated inference can be wrong even when the underlying data points are real.

When we think of 'personal data', we usually picture what we voluntarily type into a form field: our name, email address, or phone number. That is only the first layer, known as provided or declared data.

While you use an app or browse a website, the service also logs observed data: what time you connect, from which device model, how long you pause on a screen, what you search for, or your approximate location.

Combining those two layers, many systems compute a third layer: inferred data. An inference is a statistical deduction or prediction about you—for example, estimating that you have young children, are facing financial stress, or lean toward a political view. Even when built on real traces, an inference is an automated hypothesis: it can be completely wrong if you searched for a gift for someone else, share a device at home, or looked up a one-off question.

Technical detail: how an observed trace turns into a profile

In data architecture, usage events (clicks, dwell time, session identifiers, or browser metadata) are aggregated into activity logs. Profiling systems compare those logs against patterns from other users to assign probability labels or commercial segments.

The practical privacy problem is twofold: first, you rarely see which inferred labels were stored on your account; second, those labels can shape which ads, prices, recommendations, or filters you receive, even when the original deduction was a false positive.

Worked example: the fictional 'Neighborhood Market' app

Mateo installs a local delivery app to order diapers and a thermometer at 2:15 a.m. because he is helping his sister with her newborn baby.

1. Provided (declared) data

Name 'Mateo R.', email address, and street address typed when creating the account.

→ Mateo explicitly handed these over to receive the order.

2. Observed data

Connection at 2:15 a.m., Android phone at 14% battery, and three consecutive searches in the infant pharmacy section.

→ The app recorded this automatically from his interaction and device.

3. Inferred data (and why it fails)

Label assigned to his profile: 'New parent with urgent late-night purchase; high sensitivity to baby subscription offers'.

→ The system turned a one-time purchase for his nephew into a permanent—and mistaken—household attribute about Mateo.

Distinguishing the three layers helps you see what you control when filling out a form (provided data), what you generate by interacting (observed data), and what the system assumes without asking you (inferred data).

Decision for Module 1: In the fictional app 'Routes & Reads', Lucía uses her account to look up three children’s storybooks as a gift for her godson and downloads an offline map. The system logs: (A) her email lucia@example.test, (B) the minutes she spent reading each book synopsis at 11:30 p.m., and (C) the commercial label 'Household with preschool-age children'. Which evaluation accurately classifies all three elements?

Choose the option that best distinguishes the origin of each data point and the nature of the final conclusion:

Optional real-world action · Apply what you learned today

Review the ad topics or inferred interests a platform has assigned to your account

  1. Open the privacy or ad settings of an account you use frequently (for example, in Google Account: Data & privacy > My Ad Center; or on a social network: Settings > Ad preferences).
  2. Inspect the list of topics, age brackets, interests, or commercial categories the system has deduced from your activity.
  3. Turn off ad personalization or remove inferred categories you do not want used to profile you.
Verification check: Reload the ad preferences panel and confirm that personalized ads show as 'Off' or that the inferred categories are no longer active.
Main tradeoff or limit: Turning off ad personalization does not reduce the number of ads you see or stop all operational data collection inside the service, but it halts the use of that inferred profile to target you.
Module 2 of5 min read + interactive exercise

Module 2. Privacy and security: what protects your accounts and what the service still sees

If you protect an account with a unique password, passkey, or second factor, does that mean your activity inside that service is now private?

Observable objective:: Explain which threats a password manager, a passkey, and a second factor stop, what data an authorized service can still collect, and why account recovery is an essential part of the action.

Digital security and digital privacy need each other, but they answer different questions. Security protects your accounts and devices against unauthorized access, theft, or tampering. Privacy additionally asks what data the service you legitimately signed into collects, what it uses that data for, and who it shares it with.

Each sign-in tool covers a specific security risk: a password manager lets you use a long, unique password on every site so a breach on one website does not unlock the others; a passkey or two-factor authentication (2FA) adds a device-bound proof to stop an attacker even if they stole your password.

However, none of those defenses stops the authorized service itself from logging your purchase history, internal searches, or granted permissions. Moreover, hardening sign-in without preparing account recovery (backup codes and an up-to-date recovery email or phone) creates another real hazard: locking yourself out of your own account if your phone is lost or damaged.

Technical detail: why passkeys and authenticator apps resist phishing better

A traditional password is a shared secret: if you type it into a fake page impersonating your bank or email (phishing), an attacker can copy and use it immediately. SMS codes add a useful barrier against leaked passwords, but they can be intercepted through SIM swapping or typed into a phishing site.

An authenticator app (TOTP) generates temporary codes locally on your phone without relying on cellular networks, and a passkey (built on WebAuthn/FIDO2) uses public-key cryptography bound to the website’s real domain: because the key only responds to the authentic domain, it will not fire on an impostor page.

Worked example: Sofía secures her account on the 'OpenClassroom' platform

Sofía uses an online course platform. She enables a passkey, saves a unique password in her manager, and writes down her recovery codes, but leaves all commercial history and partner-profiling toggles enabled.

What IS protected (Security against intruders)

If a password Sofía used on an old forum leaks, an outside attacker cannot break into her 'OpenClassroom' account.

→ A unique password and passkey block unauthorized access.

What is also covered (Continuity and recovery)

If Sofía’s phone breaks, she can use her backup codes stored safely at home to regain access without losing her certificates.

→ Practical security includes planning how you will sign back in when your primary device fails.

What the authorized service CAN STILL collect (Privacy)

'OpenClassroom' still knows which courses Sofía opens, what time she connects, what city she connects from, and which interests to share with ad partners unless she adjusts her privacy settings.

→ Authenticating securely does not limit what the authorized platform collects by design.

You need both halves: access and recovery controls so strangers cannot enter your account, and minimization and privacy settings to limit what your account hands over to the provider itself.

Decision for Module 2: Carlos sets up a unique 20-character password with a password manager on his 'HomeStore' shopping account and enables two-factor authentication on his phone, but he does not save recovery codes and agrees to share his purchase history with partner companies for 'personalized offers'. Which diagnosis accurately describes his situation?

Compare which risks Carlos reduced and which remain open:

Optional real-world action · Apply what you learned today

Check your sign-in method and recovery plan on one important account

  1. Pick your primary email account or the account where you store personal documents and open its Security section.
  2. Verify that you have a passkey or two-factor authentication enabled, and generate or review your one-time backup recovery codes.
  3. Store those recovery codes in a safe place separate from your phone (for example, printed under lock and key or in a password manager accessible from another device), then open the Privacy section of the same account to turn off commercial history or ad tracking you do not need.
Verification check: Confirm you can answer yes to two questions: 'Would an intruder who only has my password hit a second factor or passkey wall?' and 'If I lose my phone today, do I know exactly where my backup codes are to recover my account?'
Main tradeoff or limit: Storing recovery codes requires keeping them safe from others, and signing in on a new device will take a few extra seconds when confirming the second factor.
Module 3 of4 min read + prioritization exercise

Module 3. What to protect first: prioritizing by consequences and effort

When you have an email account, shared cloud documents, and social profiles, where should you start so you do not get overwhelmed?

Observable objective:: Prioritize the first privacy and security change across a primary email account, a shared cloud document, and a social profile by weighing likely consequences and required effort.

Trying to fix every digital setting in a single afternoon often leads to fatigue and abandonment. Basic threat modeling avoids that trap by asking four simple questions: what do I want to protect, from whom, what are the likely consequences if it fails, and how much effort does each change take?

Not all accounts carry the same weight. Your primary email account is usually the central node: if someone takes control of your email, they can request password resets for your cloud storage, social networks, and financial services. Similarly, a folder or document containing IDs or contracts set to 'Anyone with the link' exposes sensitive data right now and takes thirty seconds to restrict.

Prioritizing means starting with changes that combine high impact (preventing chain-reaction account takeovers or closing active leaks of sensitive data) with low or moderate effort, before spending hours fine-tuning secondary accounts.

Technical detail: cascading dependencies between accounts and single points of failure

Most consumer online services rely on email-based recovery flows (reset links or verification codes). In security architecture, that makes your primary email a single point of failure: its compromise cascades downward to every service tied to that address.

Evaluating impact × effort lets you resolve single points of failure and active public exposures (such as open cloud links) first, capturing the largest risk reduction in the first ten minutes.

Worked example: three pending fixes on Daniel’s devices

Daniel has 10 minutes before leaving home and notices three issues: (1) his primary email uses a reused password without 2FA; (2) a spreadsheet with family tax data is shared as 'Anyone with the link'; and (3) an old hobby forum profile shows the city where he lived five years ago.

Priority 1 (Immediate · 1 minute): Restrict the tax spreadsheet link

Change access from 'Anyone with the link' to 'Restricted' in his cloud drive.

→ Very high impact (closes an active exposure of tax and ID data) and minimal effort (three clicks).

Priority 2 (Today · 8 minutes): Harden primary email + save backup codes

Enable a unique password/passkey and 2FA on his primary email, storing recovery codes.

→ Structural impact: protects the master key that controls recovery for all his other accounts.

Priority 3 (When time allows): Clean up the old hobby forum profile

Log into the old forum and remove the outdated city field or delete the account.

→ Low consequence compared to the open tax file and unprotected primary email.

Prioritizing is not ignoring minor details forever; it is making sure minor details do not distract you while your master key or a sensitive document stays wide open.

Decision for Module 3: Review Elena’s fictional situation: (1) Her primary email account (used to reset her bank and cloud storage) has no second factor and reuses a password from a leaked site; (2) A cloud document containing a scanned copy of her lease and ID is set to 'Anyone with the link can view'; (3) Her public social profile shows her favorite bands and birth month. If Elena has 15 minutes today, which order of action makes the most sense by consequences and effort?

Pick the sequence that addresses the most severe and immediate risks first:

Optional real-world action · Apply what you learned today

Make your own 3-line priority list before changing settings at random

  1. Write down on a sticky note or blank file: (1) Which email account can reset your most important accounts; (2) Which cloud folder or file holds sensitive documents (IDs, contracts, health, or finances); (3) Which social profile or public account exposes more than you want.
  2. Check the sharing status of that sensitive file or folder (1 minute) and the sign-in + recovery setup of that primary email (8 minutes).
  3. Leave secondary accounts or old forums for a separate session once those first two items are verified.
Verification check: You have a clear order of action and have verified that neither your primary email nor your most sensitive shared document remains exposed.
Main tradeoff or limit: Accepting that you will not clean up every old account in a single day feels less 'complete' at first, but it ensures you actually finish the changes that protect you most.
Module 4 of5 min read + interactive permissions panel

Module 4. Permissions and minimization: grant only what the current task requires

Why does a simple notes app ask for your microphone, contacts, and precise location—and what changes when you switch from typing text to recording a voice note?

Observable objective:: Choose appropriate permissions for a fictional notes app when writing a text list, retry when the task switches to recording a voice note, and identify which decision changes and which stay denied.

Data minimization is a simple rule of thumb: grant only the access required for the specific task you want to perform right now, and for the shortest time needed.

Many mobile apps request access to the microphone, camera, contacts, photo library, or precise location the first time you open them, even if those sensors are only used for optional features you may never touch. Granting 'Allow all the time' turns an occasional convenience into continuous exposure.

Modern mobile operating systems let you decide with granularity: you can deny permissions unconnected to your task, grant access 'Only while using the app' or 'Ask every time', give approximate instead of precise location, or select individual photos instead of opening your entire gallery.

Technical detail: runtime permissions, background access, and third-party SDKs

When you grant a permission to an app, every software library (SDK) bundled inside that app—including analytics or advertising modules—can potentially access the data allowed by that permission unless the operating system isolates it.

Restricting a permission to 'Only while using the app' or 'Ask every time' prevents background collection when the screen is off or when you are using other apps, and revoking Contacts prevents exposing third parties stored in your address book.

Worked example: what each sensor exposes in 'ClearNotes'

Before using the interactive simulator below, compare what each permission actually unlocks on your phone:

Microphone permission

Allows the app to capture audio from your surroundings whenever the permission is active.

→ Unnecessary for typing text; necessary only while actively recording a voice note.

Contacts permission

Allows reading names, phone numbers, emails, and notes across your entire address book, including people who never agreed to use that app.

→ Unnecessary both for writing a personal list and for recording a personal voice note.

Precise location permission

Hands over your exact coordinates (street and building) derived from GPS, Wi-Fi networks, and nearby cell towers.

→ Unnecessary for taking notes or recording audio, unless you explicitly want to attach a map to a specific entry.

Denying a permission today does not break the app: if you tap the voice recording button later on, the system will prompt you at that moment.

Permissions simulator: fictional app “ClearNotes”

Active situation: You just installed the fictional app 'ClearNotes' to type a grocery list in plain text ('rice, coffee, soap') and check it on your phone. When you open 'ClearNotes' for the first time, three permission prompts appear. Choose which level you would grant to each one to write a text list:

Permission: Microphone

“To record voice notes and dictate reminders”

Recommended setting (Don’t allow (or Ask every time)): Typing text with your keyboard does not require capturing audio. Keeping the microphone closed prevents accidental recordings.

Permission: Contacts

“To find friends who also use ClearNotes”

Recommended setting (Don’t allow): A personal shopping list does not need to upload the names, phone numbers, and emails of other people stored in your address book.

Permission: Precise location

“To attach the exact place where you wrote each note”

Recommended setting (Don’t allow): Jotting down what to buy does not require logging your GPS coordinates or mapping your daily routines.

When the task is writing a text note for yourself, none of the three permissions is necessary. Denying them applies data minimization without losing the core function.

Decision for Module 4: In the 'ClearNotes' simulator: first configure permissions for writing a text shopping list, then switch to the second attempt to record a voice note.

Observe how the microphone decision changes when the task changes, while contacts and location remain denied:

Optional real-world action · Apply what you learned today

Audit Location, Microphone, and Contacts permissions on your phone

  1. Open your phone’s Permission Manager (Settings > Security & privacy > Privacy > Permission manager on Android; Settings > Privacy & Security on iPhone).
  2. Open 'Location' first and check which apps are set to 'Allow all the time' (or 'Always'). Switch any app that does not need continuous tracking to 'While using the app' or 'Never / Don’t allow'.
  3. Next, open 'Microphone' and 'Contacts' and remove access from apps where you do not use voice features or address-book syncing.
Verification check: Confirm that the 'Allow all the time' Location list contains only apps whose background function you consciously chose (or is completely empty).
Main tradeoff or limit: If you later use a secondary feature that needed that permission (such as sending a voice note or sharing live location), your phone will ask for authorization right then—confirming why it needs it.
Module 5 of6 min read + diagram + interactive exercise

Module 5. Who can see what: trackers, HTTPS, VPNs, and end-to-end encryption

When you turn on HTTPS, a tracker blocker, a VPN, or an end-to-end encrypted chat, who stops seeing your data and who still observes part of it?

Observable objective:: Identify which information (content, visited domain, IP address, cross-site activity, or metadata) remains visible to the local network/ISP, the VPN provider, the destination service, and third-party trackers in each scenario.

One of the most common misunderstandings online is looking for a single magic tool that 'hides everything from everyone'. In practice, each protection covers a different leg of communication and shields against a different observer.

HTTPS encrypts the connection between your browser and the website so that neither a coffee-shop Wi-Fi network nor your internet service provider (ISP) can read the exact page you view or what you type into a form; however, your ISP usually still sees which general domain you connected to, and the destination website receives everything you do inside it.

Tracker blockers stop external companies embedded across many pages from linking your visits across different websites. A VPN hides which domains you visit from the local Wi-Fi and your ISP, but shifts that network visibility to the VPN provider and does not hide your identity from a website where you sign in. Finally, end-to-end encryption (E2EE) prevents the messaging provider itself from reading the content of your messages, though the provider still sees operational metadata such as who communicates with whom and when.

Technical detail: the boundary between encrypted content and communication metadata

Whenever you send a message or load a webpage, two kinds of information travel: content (the text of the message, the photo, or the search terms inside the site) and metadata or routing data (source and destination IP addresses, the server name in DNS or TLS SNI lookups, file size, and timestamps).

That is why a tool that encrypts content (such as HTTPS in transit or end-to-end encryption in messaging) still leaves operational signals needed to deliver packets. Understanding that boundary prevents expecting a VPN to do the job of end-to-end encryption, or expecting HTTPS to stop tracking cookies inside a site.

Worked example: checking medical results on coffee-shop Wi-Fi

Elena is connected to coffee-shop Wi-Fi. First she reads articles about a diagnosis on an HTTPS health portal, and then she sends a photo of her lab results to her doctor using an end-to-end encrypted messaging app.

What the coffee-shop Wi-Fi / ISP sees

Sees that Elena’s phone connected to the general domain of the health portal and to the messaging app’s server, but CANNOT see which article she read or open the photo of her lab results.

→ HTTPS and end-to-end encryption protect content against the local network.

What changes if Elena turns on a trustworthy VPN

The coffee shop and ISP stop seeing even the name of the health portal (they only see encrypted traffic to the VPN), but now the VPN provider sees that destination domain.

→ A VPN shifts network observation from the local Wi-Fi/ISP to the VPN server.

What the E2EE messaging provider sees

Cannot read the text or view the lab photo, but does log that Elena’s account sent a file to her doctor’s account at 4:40 p.m.

→ End-to-end encryption protects content, not all metadata or the recipient’s unlocked screen.

Choosing well starts with asking: from which observer do I want to protect which piece of data right now (the local network, cross-site trackers, or the messaging server)?

Comparison diagram: who still observes what in each scenario?

Follow the path from your device (Browser / App) → Local network & ISP → VPN provider (if used) → Destination service & Third-party trackers:

Visibility comparison by observer depending on the tool used
Scenario and actionLocal Wi-Fi / ISPVPN providerDestination serviceThird-party trackersMain limit
1. Standard web browsing with third-party trackers (over HTTPS)
You visit several news sites and online stores using a default browser without tracker blocking.
Partial / Metadata
Sees your IP address, connection time, and the general domain you visit (e.g., news.example), but not the exact article path or form inputs thanks to HTTPS.
Not involved
Not involved in this scenario.
Visible
Sees your IP address, the exact pages you open, the forms you submit, and your account if you are signed in.
Visible
Receive requests from every site where they are embedded and can link your visits across different websites to build a commercial profile.
HTTPS encrypts network transit, but it does not stop the website itself or embedded third-party trackers from logging your activity.
2. Browsing with HTTPS + enhanced tracker protection enabled
You visit the same website using HTTPS and a browser that blocks cross-site cookies and known trackers.
Partial / Metadata
Still sees that you connect to the site’s domain and the approximate data volume, but cannot read or alter content in transit.
Not involved
Not involved in this scenario.
Visible
Still sees everything you do inside its own website and your connecting IP address.
Hidden / Encrypted
Their reach drops sharply: the browser blocks known tracking scripts or isolates their cookies so they cannot link your activity across sites.
Blocking third-party trackers curbs cross-site profiling, but the first party (the website you visit) still sees what you do on it, and your ISP still sees the domain.
3. Connecting through a VPN + HTTPS
You turn on a trustworthy VPN before opening an HTTPS website on a shared Wi-Fi network.
Hidden / Encrypted
Only sees that your device maintains an encrypted tunnel to the VPN server; it no longer sees which web domains you look up.
Partial / Metadata
Receives your connection: knows your real IP address and sees which domains or servers you connect to, though it cannot read internal content protected by HTTPS.
Visible
Sees the VPN’s IP address instead of yours, but if you sign in with your account it knows exactly who you are and what you do on the site.
Partial / Metadata
Do not see your home IP address, but if you do not block trackers they can still recognize you via cookies, active logins, or browser fingerprinting.
A VPN shifts trust from the local network or ISP to the VPN provider; it does not make you anonymous when signed in, nor does it replace tracker blocking.
4. End-to-end encrypted (E2EE) messaging
You send a message or document to another person using an end-to-end encrypted messaging app.
Partial / Metadata
Sees that your device communicates with the messaging app’s servers, but cannot read the message or see who you are talking to.
Not involved
If you use a VPN, it hides from the local network that you are using that messaging app, but does not change who can read the message.
Partial / Metadata
The messaging provider CANNOT read the encrypted text, audio, or files, but it does see operational metadata (which accounts or phone numbers communicate, when, and how often).
Hidden / Encrypted
Have no access to the encrypted content of the conversation.
End-to-end encryption protects message content from intermediaries, but metadata still exists, the recipient receives the full message, and unencrypted cloud backups or an unlocked phone can expose it.
Decision for Module 5: Andrés uses his home browser to check three different online stores and then sends a private document to a colleague via an end-to-end encrypted messaging app. Which statement accurately describes who can still see what?

Review the comparison matrix above and select the accurate statement:

Optional real-world action · Apply what you learned today

Enable enhanced tracking protection and HTTPS-Only mode in your browser

  1. Open the privacy and security settings of your main browser (for example, in Firefox: Settings > Privacy & Security > Enhanced Tracking Protection; in Chrome, Brave, or Safari: Privacy and security section).
  2. Confirm that cross-site tracking cookie blocking or enhanced tracking protection is enabled, and turn on 'HTTPS-Only Mode' (or 'Always use secure connections').
  3. Open a news or shopping website and click the shield or tune icon next to the address bar to see how many trackers the browser blocked and confirm the connection uses HTTPS.
Verification check: Clicking the shield or info icon next to the web address (URL) shows 'Connection secure (HTTPS)' and an active blocked-trackers indicator.
Main tradeoff or limit: On very old websites that do not support HTTPS you will see a warning before proceeding, and under the strictest blocking level an embedded form or video may occasionally require temporarily toggling the shield off for that single page.
Module 6 of6 min read + redaction and audience exercise

Module 6. Before sharing, posting, or pasting into an AI: audience, redaction, and re-identification

If you delete a person’s name before sharing a document or pasting it into an AI tool, is that enough to keep them from being identified?

Observable objective:: Adjust the audience setting of a shared document, select which fields to remove or generalize before pasting into an AI tool, and explain how remaining details can still identify someone by combination.

Every time you share a cloud document, post a screenshot, or paste text into an artificial intelligence assistant to summarize it or draft a reply, you make two privacy decisions: who can access the document and how much identifying detail the text contains.

A common mistake is assuming that deleting a first and last name makes a text 'anonymous'. In reality, people are often identifiable through a combination of details (known as quasi-identifiers): if a document mentions 'the only morning-shift accountant at the Mérida branch who requested medical leave for surgery on October 14', anyone in that environment will immediately know who it is, and the external service will receive unnecessary employment and health data.

Before clicking 'Share' or 'Send', apply two filters: first, restrict the file’s audience to those who genuinely need to read it; second, if you use an external AI tool, strip direct identifiers (names, ID numbers, phone numbers, emails) and generalize unique details so only the structure of the problem remains.

Technical detail: why re-identification happens through combined attributes

Data protection distinguishes between direct identifiers (an ID or passport number, a personal phone number) and quasi-identifiers (job title, postal code, exact age, event dates, or rare traits). Combining just two or three quasi-identifiers often narrows a group of thousands down to a single individual.

Furthermore, when you send text to an external AI service, the prompt leaves your device for the provider’s servers and—depending on your account type and settings—may be retained in history logs or reviewed to train models. Redacting locally on your device before pasting is the safest barrier.

Worked example: from a real email with third-party data to a clean AI prompt

A coordinator wants to ask an AI assistant to draft a polite reply to a workplace request:

Original text (excessive)

“Draft a reply for Jorge Pineda (ID 15.220.419, phone 600-123-890), the sole refrigeration technician at our Valdivia branch, who is asking to work remotely for two weeks after his eye surgery on Monday the 3rd at Austral Clinic.”

→ Exposes another person’s name, ID, phone number, unique job role, city, and medical diagnosis.

Incomplete attempt (removes only the name)

“Draft a reply for the sole refrigeration technician at our Valdivia branch (ID 15.220.419), who is asking to work remotely for two weeks after his eye surgery on Monday the 3rd at Austral Clinic.”

→ Still immediately identifiable via the ID number, unique role, city, and medical date.

Minimized and generalized version (recommended)

“Draft a cordial reply confirming a temporary two-week task adjustment for a technical team member for justified health reasons, explaining how on-site shifts will be covered.”

→ The AI gets all the context needed to write a great template without receiving any personal or medical data.

Once the AI returns the general draft, you fill in the recipient’s name locally in your email client before sending.

Redaction & audience simulator: building notice before sharing or using AI

Step A. Choose the audience setting for the shared document

Recommended: limits the audience of the original document to those managing the repair and prevents accidental edits.

Step B. Check which segments you should remove or generalize before pasting the text into an AI assistant

Full name and ID number directly identify the person and add nothing to help an AI summarize the issue or polish the wording.

Including real phone numbers or emails in an open link or an AI prompt exposes contact details unnecessarily.

Even if you remove her name, saying 'the only pediatrician on floor 3B who had surgery on October 4' immediately identifies Valeria to any neighbor and hands unnecessary health data to third parties.

This passage contains the facts needed to draft the notice or summarize the proposal without identifying a specific individual.

Live preview of the text you would send to the AI assistant

«Please draft a formal summary of this case: Requester: [Building resident] Contact: [Internal building management channel] Context: “A resident with temporary reduced mobility requires advance notice before elevator service is interrupted” Issue: “Water leak in the shared hallway for the past ten days and request to schedule elevator repairs with 48 hours’ advance notice while comparing three vendor quotes” »

Decision for Module 6: In the interactive exercise on Valeria’s building notice (right above), after choosing the document audience and the fields to remove or generalize before pasting into an AI tool, which principle best summarizes the right approach?

You can test the interactive audience and field-redaction panel first or answer directly:

Optional real-world action · Apply what you learned today

Use the 'clean copy' habit and review your AI assistant’s privacy settings

  1. Before pasting an email, contract, or table into an AI assistant, copy it into a blank note first and replace names, ID numbers, phone numbers, and unique roles with generic placeholders like [Client A], [City], or [Health reason].
  2. Open the settings of the AI tool you use (for example, Data Controls / Privacy in your assistant account) and check whether your conversations are used for model training.
  3. Turn off model training on your chats if you prefer to limit reuse of your prompts, or use temporary chat modes when available.
Verification check: Read the text you are about to send and run this quick test: 'If a coworker or neighbor read this paragraph without the name, could they still guess who it is from the combination of details?'
Main tradeoff or limit: Replacing specific details before querying an AI and adding the name back locally takes an extra minute or two, but prevents handing over identifiable data about yourself or others.

Next practical step · Interactive simulator

Put your decisions to the test in the digital privacy lab

In the practice route you will find four interactive stations with feedback across four visible criteria, optional hints, a second attempt that changes each situation, and a final capstone scenario on organizing an event without exposing participants.

3. Curated & verified directory · Full reading experience without JavaScript

Curated directory of websites and practical guides to learn digital privacy

These portals, manuals, and help centers were editorially verified on November 12, 2026 They are grouped by the practical task you want to accomplish today, showing the publisher, intended audience, and jurisdiction note whenever legal frameworks are referenced.

1. Getting started and reducing everyday data exposure

Beginner-friendly guides to review daily habits, clean up digital traces, and run a practical self-assessment without jargon.

Tactical Tech

Data Detox Kit

What it helps you do: Short, everyday activities to tidy up phone traces, adjust app settings, reduce personal data exposure, and build healthier digital habits without overwhelm.

Recommended audience: Beginners and general readers looking for bite-sized exercises that take only a few minutes.

Verified: November 12, 2026Open official guide ↗

2. Configuring browsers, mobile phones, and social networks

Practical walkthroughs for limiting cross-site tracking and tightening account visibility.

Spanish Data Protection Agency (AEPD)

Protect your privacy (practical guides and walkthroughs, Spanish)

What it helps you do: Step-by-step walkthroughs for reviewing privacy settings in browsers, mobile operating systems, apps, and social networks, plus guidance for families and minors.

Recommended audience: General readers looking for visual checklists and settings walkthroughs for popular services.

Jurisdiction note: Technical configuration steps apply in any country, but legal references, procedural rights, and complaint channels correspond to Spain and the European Union (GDPR and LOPDGDD).

Verified: November 12, 2026Open official guide ↗

Mozilla Support (Official Firefox Documentation)

Enhanced Tracking Protection in Firefox for desktop

What it helps you do: Understand and configure Standard, Strict, and Custom levels to block cross-site tracking cookies, social trackers, and fingerprinters, and learn how to troubleshoot if a site breaks.

Recommended audience: Beginners and intermediate users who want official instructions for hardening their desktop browser.

Verified: November 12, 2026Open official guide ↗

3. Threat modeling, metadata, and higher-risk contexts

Reference manuals for building a threat model, understanding browser fingerprinting and network metadata, or strengthening defenses in journalism and civil society.

Electronic Frontier Foundation (EFF)

Surveillance Self-Defense (SSD)

What it helps you do: Learn how to build a threat model, choose encrypted communication channels, protect devices against physical or remote intrusion, and follow security playlists tailored to different roles.

Recommended audience: From beginners who want to understand why each tool works to journalists, researchers, and human rights defenders.

Verified: November 12, 2026Open official guide ↗

Tor Project

Tor Community Digital Security & Training Resources

What it helps you do: Dive deeper into browser fingerprinting, connection metadata, identity separation, and how the Tor network works—and where its limits lie.

Recommended audience: Readers looking for deeper coverage of network privacy, technical anonymity, and the limits of conventional browsing.

Verified: November 12, 2026Open official guide ↗

Amnesty International Security Lab

Digital Security Resource Hub

What it helps you do: Guides on hardening accounts and mobile phones, defending against targeted phishing or spyware, and practical advice for organizations in higher-exposure contexts.

Recommended audience: Journalists, civil society members, human rights defenders, and users in higher-risk environments.

Verified: November 12, 2026Open official guide ↗

4. Primary platform documentation (exact menu paths and versions)

Menu names can vary by phone manufacturer or operating system version. Whenever a path differs on your device, look for the functional goal described below or check the manufacturer’s official help page:

Google Android

Change app permissions on your Android phone

Functional goal to look for: Locate the Permission Manager by category (Location, Microphone, Contacts, Camera) and enable automatic permission reset for unused apps.

If the menu varies on your version: The exact menu path may be called 'Security & privacy > Privacy > Permission manager' or 'Apps > Permissions' depending on your Android version and device manufacturer.

Verified: November 12, 2026Official manual ↗

Apple iPhone / iOS

Control access to information in apps on iPhone

Functional goal to look for: Check Settings > Privacy & Security to see which apps access location, microphone, contacts, or photos, and turn off Precise Location when approximate location is enough.

If the menu varies on your version: On iOS 17 and later, you can grant access only to selected photos or specific contacts without opening your entire gallery or address book.

Verified: November 12, 2026Official manual ↗

Google Account

Turn on 2-Step Verification and recovery methods

Functional goal to look for: Set up passkeys or a second factor and download one-time backup codes so you do not lose access if your phone is lost or replaced.

If the menu varies on your version: On other providers (Microsoft, Apple Account, Proton, Yahoo), look under 'Security' > 'Two-step verification / Multi-factor authentication' and 'Recovery codes'.

Verified: November 12, 2026Official manual ↗

Google Drive

Share files and folders and manage access

Functional goal to look for: Change open links ('Anyone with the link') to 'Restricted' and adjust roles between Viewer, Commenter, and Editor.

If the menu varies on your version: In OneDrive, Dropbox, or iCloud Drive, the equivalent control appears under 'Share' > 'Link settings' or 'Who can access'.

Verified: November 12, 2026Official manual ↗

Dedicated guides in editorial preparation (pending publication verification)

On Libertatis we only link to published, verified articles. The following six complementary topics are part of our practical coverage in the modules above and are tracked as editorial candidates pending individual publication verification:

Mobile app permissions

Dedicated guide on runtime permissions, one-time access, and auditing sensors on Android and iOS.

Editorial candidate pending individual publication verification

Mobile phone privacy

Practical walkthrough on advertising IDs, lock-screen notifications, cloud backups, and profile separation on mobile devices.

Editorial candidate pending individual publication verification

WhatsApp privacy and security

Configuring two-step verification (PIN), end-to-end encrypted cloud backups, profile visibility, and metadata limits.

Editorial candidate pending individual publication verification

Password managers

How password managers work, what they solve against password reuse, and how to protect your master vault.

Editorial candidate pending individual publication verification

Two-factor authentication (2FA) and passkeys

Comparison between SMS codes, authenticator apps (TOTP), passkeys, and hardware security keys.

Editorial candidate pending individual publication verification

Tracker blockers

Differences between built-in browser tracking protection, content-blocking extensions, and limits against fingerprinting.

Editorial candidate pending individual publication verification