Android's photo permission isn't consent, and Italy just priced that in
A €805M fine on Meta shows that tapping "allow" on a vague prompt isn't a real choice. Google and Meta should make gallery access visible and opt-in.
AI-generated image
How many times have you tapped "Allow" on a permission prompt without reading it? That reflex is exactly what Italy's data protection authority just put a price on: €805 million.
Facebook asked once, for everything, and called it consent
According to the Italian authority's decision, as reported by Quotidiano Nazionale, Facebook for Android accessed users' photo galleries through a feature called "Get suggestions from the gallery when you browse Facebook". It proposes photos and videos from your phone as things to post. The regulator's complaint is about how permission was obtained. Facebook used a single authorization for the whole app, covering several features with different needs. The prompt carried a generic notice about enabling "other features". And the feature was on by default. The regulator estimates about 20 million Android users in Italy are affected.
Meta disagrees. Its spokesperson called the fine disproportionate and said users must grant a permission on their device before the feature can work, and can change it at any time in settings. The company plans to appeal.
"The user pressed Allow" is a weak defense
Meta's argument is the one the whole industry leans on: the operating system asked, the user said yes. The regulator's reasoning, as Hardware Upgrade describes it, is that a blanket permission for the whole app can't be read as consent to one specific purpose. Think about what that permission can mean in practice. Per the regulator, a full grant can let the app read the entire photo archive, usually permanently, and neither the user nor the authority can verify what is actually used.
We think that is the heart of the case. A permission is a technical gate. Consent is an informed decision about a purpose. When one gate covers a dozen purposes, pressing it tells you almost nothing about what the person agreed to. The legal expert quoted by Quotidiano Nazionale, Giuseppe Vaciago, calls the key principle the "granularity of consent": you can't ask for a general, unclear, let alone automatic authorization for different activities.
The fine is big, but the real ruling is the design change
The headline number is dramatic, though in context it is about 0.47% of Meta's worldwide revenue, according to the regulator's figures. We think Meta can absorb that. What it can't easily absorb is the order attached to it. Within 60 days, Meta must ask for separate, specific consent for the gallery feature, explain what it does and how to turn it off, and keep it off by default, even for people who already granted photo access. That last point matters most: an old "yes" no longer counts as a new one.
The legal route is also notable. Meta argued the Irish authority should handle the case. The Italian regulator rejected that, because the ePrivacy rules it relied on don't have the GDPR's one-stop-shop mechanism. If that holds up on appeal, national regulators could get a sharper tool against app-level data access than many companies expected. That is our reading, not a settled outcome.
What Google and Meta should do now
Quotidiano Nazionale notes that Apple users aren't affected, since iOS has its own explicit-consent mechanism. We wouldn't over-read that. The reports we've seen don't explain how Android's own prompts work, and the case targets Facebook's conduct, not Google's. But it does hand Google an obvious design question: why should a social app be able to hold lasting access to everything in your gallery when it only needs one picture?
Our recommendation is simple. Meta should make gallery suggestions opt-in worldwide, not only where a regulator forces it, with a plain-language screen saying what is read and why. Google should make purpose-specific, visible access the easy path for photos and videos, and make broad access the exception that takes extra effort.
The investigation did not look at whether gallery content was used to train AI models, which Meta denies. The regulator says it will examine separately a cloud-processing feature launched in 2026. That is the part to watch: when features start processing your photos in the cloud, vague consent stops being a nuisance and becomes a real risk.
Verdict
For now, check your Facebook permissions and turn off photo suggestions if you never asked for them. The bigger lesson is for the industry: a buried toggle and a vague prompt aren't consent. Regulators are starting to say so with numbers that get noticed.
No topics yet: start the first one.
More stories
Hiding your WhatsApp number is overdue, but defaults will decide if it matters
WhatsApp usernames promise to keep your phone number private. The idea is right, but the real test is who gets protected by default and whether it holds up in groups.
"Don't click anything" is dead advice: patching is your real shield
A suspected Russia-linked campaign hit iPhones and Androids through ordinary websites, no clicks needed. Fully patched phones were safe, so vendors should patch for us.
WhatsApp's tappable text is handy, but your chat is becoming a launcher
Dotted-underline numbers, dates and addresses save a few taps. They may also point to Meta wanting the chat to be where things start, so we should ask what gets read.
Copilot's TV exit proves smart-screen AI is a rental, not a feature
Microsoft is reportedly pulling Copilot from Samsung and LG TVs on December 1, 2026. Anyone who counted AI as a reason to buy a screen just learned how revocable that promise is.



