Having a say in how algorithms see you, instead of accepting whatever profile they infer.
Algorithmic self-determination is your ability to decide how algorithms see you, instead of accepting the profile they build from everything you do. In practice it starts with privacy: making the data collected about you say less about you, so it is worth less to the advertisers and AI agents that rely on it.
Every search, video and product page you open becomes a signal. Platforms turn those signals into a picture of who you are, then use that picture to pick your next ad, your next video and your next "recommended for you." Most people never get a say in that picture. Self-determination means you take a hand in it.
Your profile is built for you, not by you. Platforms infer interests, habits and life circumstances from what you do, and they keep refining that picture for as long as you use them.
It doesn't stay in one place. Ad platforms group you into interest and audience categories, and data brokers buy, combine and resell behavioral data across companies. What you looked at in one place can shape what you're shown somewhere else.
More and more of that data is also read by AI. Recommendation systems, ad systems and AI agents that act or answer on your behalf all work from what's known about you, and the more precise the profile, the more it reveals.
And you mostly can't see it. Google shows some of what it uses in My Ad Center, and a few other platforms have similar pages, but these views are partial and editing them is limited. The full picture stays on their side.
There's no single switch. Here are the practical options, roughly from the most manual to the most automated.
Search and watch deliberately. On YouTube, use "Not interested" and "Don't recommend channel," and remove individual videos from your watch history. Follow topics you want more of.
Good: free, precise, works on almost every platform. Limits: it takes steady effort, and a single late-night rabbit hole can undo a week of care.
You can turn off or auto-delete Google's Web & App Activity and pause YouTube watch and search history. Auto-delete after 3 months is a good middle ground.
Good: less gets stored going forward. Limits: recommendations get worse or go generic, some features stop working, and it doesn't stop every kind of collection. It also stops your feed from learning anything new, including things you'd want it to learn.
Google's My Ad Center lets you see some of the topics and brands it associates with you, remove ones you don't want, and turn personalized ads off. Meta and others have similar controls.
Good: official, quick, reversible. Limits: turning personalized ads off changes which ads you see, not whether data is collected. The lists are partial, and none of this touches your content recommendations.
Tracker blockers and privacy-focused browsers cut down third-party tracking, and VPNs keep your IP address from the sites you visit.
Good: reduces future collection across the web. Limits: they can't change what's already been built, and they don't stop a platform from learning what you do while you're signed in to it.
Instead of subtracting, you add. Ordinary browsing on topics that aren't yours, like a home chef's recipe searches or a traveler's destination videos, gets mixed into your history. You stay yourself. Your real interests just become one part of a bigger mix, the way a drop of paint changes a glass of water.
Good: it works on the profile that already exists, and you keep personalized recommendations, they just get broader. Limits: doing it by hand is slow and repetitive, which is exactly the part worth automating.
Changing how you're seen also changes what you're shown. Once a new topic is in the mix, your YouTube home feed may start showing fly fishing or bread baking next to the usual, which breaks the recommendation loop that keeps serving you more of the same. That's a nice side effect, and many people enjoy it, but here it comes along with the privacy work rather than leading it.
MirrorMask is a native macOS app that automates the fifth option. It runs on macOS 14 or later, on Apple silicon and Intel Macs. The current version is 1.1.0.
If you want to see what this looks like on fresh accounts, we ran a 5-day experiment with a control account.
Algorithmic self-determination is your ability to decide how algorithms see you, instead of accepting the profile they build from everything you do. It starts with privacy, and it also makes your data worth less to advertisers and AI agents. You can work toward it by hand, with platform settings, or with a tool that adds ordinary browsing signals for you.
Algorithmic curation is about what you get shown: shaping your feeds so they bring you fresher, more varied things. Algorithmic self-determination is about how you are seen: having a say in the profile platforms keep about you, mainly for privacy. The two overlap, because changing your profile changes your feeds too, but curation is a side benefit here, not the goal.
It can make the profile they draw on less accurate. If an AI system builds a picture of you from your search, browsing and viewing history, a history that mixes your real interests with a persona's gives it a blurrier picture to work from. Nobody can guarantee how any particular system uses data, and this doesn't remove what has already been collected, so treat it as one layer of privacy, not a promise.
They are related, but not the same. Data poisoning usually means corrupting a system's training data. Algorithmic self-determination is about your own profile, shaped with ordinary browsing, like the searches and videos a real person makes. It changes what platforms believe about you, not how their systems work.
Nobody can promise how a platform will act, but adding signals to your own profile uses the same actions any person takes: searching, visiting pages, clicking, scrolling and watching videos. MirrorMask never posts, comments, likes or votes, and it runs at most one automatic session a day.
No. Ad blockers and VPNs reduce what gets collected from now on. They don't change the profile a platform has already built, and they don't stop a platform from learning what you do while you're signed in to it. Algorithmic self-determination works on the profile itself. The two approaches work well together.
MirrorMask is a Mac app that mixes a persona's everyday browsing into yours, up to one session a day. Free for 7 days with no card, then a one-time $39 for Pro.
Download for Mac