ainotis Join
My notis

Privacy

If you read ai notis without joining or signing up for the morning email, we keep nothing about you. If you join, we keep your email and what you choose to save, and you can download it or delete it at any time. If you sign up for the morning email, we keep your email address and when you asked for it, until you unsubscribe (the details are under The morning email).

Reading without joining

We count that a page was read on a day: a date, a page address and a number, and where the reader came from, as one of a few labels (below). For a story we also count how long it was read and how far down, in rough bands, under "What we count". Not who read it. Visits from robots (search engines, AI crawlers, feed readers) are counted apart from readers'. We tell them apart by what the browser says it is, and by whether the request carries the language and request-type lines every browser sends; we look at these while counting and never keep them. To try a new robot test before we use it, we also count, per day, which of the robot tests we are trying a reader's visit would have failed, by kind of page and as one of three words for where it came from; never the page's address, the browser's name or version, or what any line it sent said. A browser that asks not to be tracked is left out. Our own checks of the site are not counted at all: our tools name themselves, and we look at the address of each visit only to tell whether it comes from our own server, and never keep it. No cookie, no address, no third party. Reactions and case follows are counted the same way, a number per story or case and a day.

When you save a story or follow a case without joining, it is kept in your own browser and never reaches us. A cleared browser loses it. The same goes for "Not now" under Get the app: your browser remembers it on that device, and it never reaches us.

Your browser also keeps a small script from us so ai notis can be installed as an app and can say so when you are offline. It stores no pages and none of your data, adds nothing to what your browser sends, and passes every page straight to us as usual.

To stop robots and floods of requests, we hold the address a request came from in the server's memory for a few minutes, and never write it down.

What we count

For every page we count one number per day. When a page is opened, your browser tells us the address of the page you came from, and what kind of browser or robot you are. We sort that into one label from a fixed list, such as "search", "AI assistant", "social", "email", "another ai notis page" or "no source given", add one to that day's count for the page, and keep nothing else. When the label is "another ai notis page", we also keep which kind of page it was (a story, the front page, a past day's page, a case, a topic, a section, Explore, the register, search or another page), never its address; that is how we see whether readers go on to a second page. A read with no source given stays "no source given": it may be an app or a typed address, and we do not guess. A site that is not on our list is counted as "other", and its name is not kept. Some links say where they came from in their own address, as ChatGPT's do; we read that the same way, against a short list, and drop it, keeping only the name of one of our own posts (below). If you put ai notis on your home screen, opening it from there is counted the same way, with one more label, "our app": it says the app was opened, not by whom. Installing it is not counted, and the app keeps no copy of any page. We never store your address, the page you came from, or your browser's details. We set no cookie and run no tracking, and no script from anyone else; the one thing our own script measures is described next. The publisher's own visits, signed in to run ai notis, are not counted at all.

On a story page we also count how long it was read and how far down. A small script on the page, ours and nothing from anyone else, adds up the time the page was on your screen and works out how far down the story you got. When you leave the page it sends one message to us with the story's address, the time as a rough band ("under 2 seconds", "2 to 10 seconds", "10 to 30 seconds", "30 to 60 seconds", "1 to 3 minutes" or "over 3 minutes") and how far you got as one of "a quarter", "half", "three quarters" or "the end". The exact time never leaves your browser. We add one to that day's count for the story and keep nothing else: no cookie, no number that identifies you, no address, no browser details, and no link to the other pages you read. Signed in or not, it is counted the same way and not tied to your account. A visit under two seconds sends a message too, in the band "under 2 seconds", so a quick look is counted like a long one. If your browser says Do Not Track or Global Privacy Control, that part of the script does not run and nothing is sent. A page your browser only fetched ahead, before you opened it (a search engine or another site can make it do that), is not counted as a read; we add one to a number per day of such pages and, for a story, one to that story's own number for the day. That tally holds the story's address, the day and a number, which is the same as our count of reads holds, and nothing about who asked. A page of ours that our own pages fetch ahead when you press a link is the visit itself, because opening it makes no second request, so it is counted as a read like any other. When your browser does say Do Not Track or Global Privacy Control, your visit is still counted as a read like anyone's, and we also add one to a number per day of story reads that carried the request, so that we can say how many readers we cannot hear back from. That number holds a day and a count, and nothing else about you or the page. The same script also says, once, that a story was read on screen: when the page has been on your screen for three seconds, or you scroll it, it sends one message with the story's address and where your browser says the page was opened from: the other site's name, sometimes with the page's path (which your browser already tells us when it opens the page), or which of our pages, and the tag on our own links, never anything after a "?" in an address and never what you typed or searched for. We turn where it was opened from into one word from a short fixed list, such as "Instagram" or "our short address", and keep only that word. We add one to that day's count for the story; this is the number we show beside a story ("read on screen"), because a robot that never puts the page on a screen never sends it. Our link-in-bio page (/links) and the front page send the same message by the same rules, so that we can see how many readers who arrive from Instagram get as far as a story. It keeps nothing more than the count of reads does: no cookie, no number that identifies you, no address, no browser details. If your browser says Do Not Track or Global Privacy Control, it is not sent either. We also count, as a number per day, how many of these messages we turned away and for which of a short list of reasons (for example, from a robot, or for a page that is not published), and how many we counted although the message lacked two of the headers a browser usually sends. That count holds the reason and the day, and nothing about the message or who sent it.

Robots

We count robots apart from readers, by company and purpose, for example "OpenAI, training" or "Anthropic, fetching a page someone asked about". To check a robot is who it says it is, we compare its internet address with the list its owner publishes. The address is held in memory for that comparison and not stored.

Headline tests

Sometimes we test a way of writing headlines. For an hour at a time the front page shows every lead's headline written one way or the other, in an order fixed before the test starts, and we count how many times the front page was shown and how many stories were opened from it in that hour. Both headlines are true and checked against the story, and stories about safety and corrected stories are never tested. We never know which reader saw which. The result is published on the numbers page.

Social networks

A link we post on a social network carries a fixed tag naming the network and the post, such as our Instagram post of 3 October, never the person who taps it; we count the arrivals per post and per day, and from the networks we take only each post's totals, never who liked or commented.

Search engines

Google and Bing tell us how often ai notis appeared in their search results and in their AI answers, and how often it was clicked there, as totals per day or per month. That is about searches, not about you, and we send them nothing about our readers.

Why we count

To know which stories are read and how people find us. Our legal basis is our legitimate interest in running the publication (GDPR Article 6(1)(f)). These daily counts are kept for about 13 months and then deleted; for page reads, only the monthly totals remain after that.

If you join ai notis

We keep:

One cookie keeps you signed in. It holds a random code, and we keep only a scrambled copy of it. It lasts 30 days from the last day you used ai notis, and signing out ends it.

The list of what you saved says what you read and cared about, so we are careful with it. It is never public and never shown to another reader. The newsroom that writes ai notis sees only totals, such as how many members saved a story, never who.

Joining, and keeping robots out

An account exists only once someone opens the link we email to the address. When you join, your browser also does a small calculation in the background, which slows robots down. You do not solve a puzzle, and no outside service is involved.

To stop anyone using our join form or our sign-up boxes to flood a stranger's inbox, no address gets more than three emails from us an hour or ten a day. To keep count we hold, for two days, how many tries came from each internet address and for each email. Both are scrambled with a secret, and the secret for internet addresses is thrown away after two days, so the counts cannot be turned back into anyone's address. After two days the counts go too.

A sign-in link works for 15 minutes, once. An address that asks to join and never opens its link is deleted after 48 hours. We keep no copy of the emails we send. Only if our mail service could not be used would an email wait on our server instead, for up to seven days, and joining is open only while it can be used, so for a member that does not happen.

The morning email

You can ask for the morning email in a box on the front page, under a story, on About, or on your own page if you have joined. You do not have to join to get it.

We keep:

Nothing else with it: not your name, not the page you signed up on, not your internet address. An address that never opens its confirmation link is deleted after 48 hours.

The list the issues are sent to is held by Brevo, the service that sends our email: it gets your address once you confirm, and sees each issue we send you. Our emails count opens and clicks, not who made them, as explained under "Who else sees it" below. We see only totals for each issue: how many were sent and delivered, opened and clicked, and how many unsubscribed or bounced. We send the email because you asked for it (GDPR article 6(1)(a), consent), and you can take that back at any time.

To leave, use the unsubscribe link at the bottom of any issue (one click), switch it off on your own page, or write to us. Leaving deletes your address from our list and from Brevo's, not just a flag on it. If you leave by the link, Brevo stops sending at once, and we delete our record and Brevo's copy the following night. Deleting your account takes you off the morning email too. So that restoring a backup can never put you back on the list, we keep a scrambled copy of your address, which cannot be turned back into it, for seven months, longer than any backup is kept.

An issue may include one sponsor, always labelled Sponsored · Advertisement. Its link goes through our own address, so we can count the clicks; the sponsor is not told who reads or who clicks.

Who else sees it

Our emails are sent by Brevo, which sees your address and the email we send. Our emails count opens and clicks: Brevo adds a small invisible image that tells it when an email is opened, and links in our newsletter pass through Brevo on the way to us. The image is fetched from Brevo like any picture, so Brevo sees the internet address and app of the device that opens the email. Brevo counts these without linking them to your email address, and we see only totals, such as how many people opened an issue. We use them to know whether our emails arrive and what readers find useful (GDPR article 6(1)(f), our legitimate interest). If you'd rather not be counted, turn off images in your mail app. The link in a sign-in email goes straight to us, never through Brevo. ai notis runs on a server we rent from Hostinger, in a data centre in Frankfurt, Germany: Hostinger holds what is on that server for us and keeps the machine running. Nobody else gets anything: no advertising networks, no analytics companies, no data brokers. An advertisement here is a box we place ourselves; it tells the advertiser nothing about you.

Taking it with you, or deleting it

On your own page you can download everything we keep about you, as a file, and delete your account. Deleting removes your data from our database straight away. It is not just hidden.

Our backups still hold a copy for a while: the nightly copies on our server for 30 days, and the encrypted copies kept elsewhere for about six months. They are used only to put the site back after a failure, and if that ever happens we erase again every account deleted since the copy was made, before the site opens.

To do that, when you delete your account we keep its number and a random code that was made for it, and nothing else: not your email, not your name. We keep them for seven months, longer than any backup, and then they go too.

Who is responsible

Karl Horpestad, a person in Norway, runs notis and is responsible for members' data.

Why we may keep it

Your email address, sign-in keys, saves, follows and reactions are kept because you asked for an account (GDPR article 6(1)(b), contract). Reading days, and the morning email if you ask for it, rest on your consent (6(1)(a)); switching either off deletes what was kept. To keep robots and floods out we hold a request's address in memory for minutes and keep scrambled counts of tries for two days, and our emails count opens and clicks without linking them to you (6(1)(f), our legitimate interest in a site and emails that work).

Asking us, and complaining

Questions about your data go to hello@ainotis.com.

Datatilsynet, the Norwegian Data Protection Authority: Postboks 458 Sentrum, 0105 Oslo, Norway. datatilsynet.no