Cookies are the main way that big tech connects your behaviours on the Internet to your personal identity. It’s how big tech knows what posts, videos, and ads to serve you next.
So why is big tech getting rid of cookies? I’ll give you a hint: it’s not because they’ve stopped tracking you… today we’re going to learn how Ai is hiding deeper. We’re going to shed light on new Ai concepts that big tech doesn’t want you to know.
The Electronic Frontier Foundation a leading non-profit defending digital privacy and free speech for over 30 years had this to say about Federated tracking: big tech “will surely tout this as a step forward for “transparency and user control,” knowing full well that the vast majority of its users will not understand how the Federated Learning of Cohorts (FLoC) works, and that very few will go out of their way to turn it off.”
I decided that would be a good challenge to explain something that is designed to be hard to understand.
I’m Dr Ed. If you’re feeling anxious around screen time, I’m here to tell you that it’s not your fault. No one could not have predicted how Ai would fundamentally change our world. Today, Ai is changing so quickly that it’s a struggle to keep up. That’s why Ai Parenting has this blogpost.
Since we are learning Ai concepts that are hard for non-techies to understand make you you leave a comment in the chat to let me know if you’re following.
There has been a lot of news regarding changes to tracking our actions on the Internet this month. Facebook bought a full page ad in the NY Times about how Apple’s new opt-in for App Tracking hurts small businesses.
Why would Apple giving you the option to opt-in to App Tracking be bad for small businessses? We browse more than a Statue of Liberty worth of posts every single day, the milliseconds that we spend on each post tells big tech about your unconscious desires. Soon apps like Facebook will need you to opt-in to tracking in order to send this information back to the Facebook Ai. As a small business owner it means that it’s harder for their ads to target specific users based on their unconscious desires.
Don’t get me wrong, this is not about Apple trying to better protect your privacy. This is not about Apple fighting the big bad Facebook from capturing your data. It’s a walled garden strategy that is going to force more advertisers to move to Apple’s own mobile ad platform rather than those of Google or Facebook. This is about Apple staking claim to the mobile advertising market that they own and control.
How is big tech reacting to this and other legal restrictions such as Europe’s Global Data Protection Regulation (GDPR)? This is where federated tracking comes in and as I mentioned at the beginning this is going to change the Internet as you know it.
This month Google announced that they would not build identifiers or tools to track users across multiple websites starting next year. Great news right?
“Instead, our web products will be powered by privacy-preserving techology which prevent individual tracking while still delivering results for advertisers and publishers,” said David Temkin, Google’s director of product management for ads privacy and trust.
Google released their alternative to cookie tracking known as the Federated Learning of Cohorts or FLoC into Chrome this month to “enable ad-targeting based on the user’s general browsing interest, without the websites knowing their exact browsing history.”
https://www.chromestatus.com/feature/5710139774468096
“Google’s FLoC is a terrible idea” wrote the Electronic Frontier Foundation (EFF) in a public blogpost. Mark my words, in several years you will yearn for the days of Cookie tracking when you could at least know what companies were tracking about you. What is happening today parallels what happened to Ai in the 1990s when people like Target dad learned things from Ai suggestions that they didn’t know about their own families. You’ll see how history is repeating itself. Big tech is stopping the creepiness but not the creeping.
We need to be curious about these three topics to understand FLoC
- Baking FLoC Cookies
- Data Rights
- Unconscious Advertising
https://www.eff.org/deeplinks/2021/03/googles-floc-terrible-idea
Baking FLoC Cookies
One way to understand federated tracking is to think about the analogy of baking cookies in 1961. Imagine this world without smartphones or the Internet.
You first got a sugar cookie recipe from page 152 of the classic red-orange Betty Crocker cookbook. Over time you found it tastes better if you add mint chocolate chips and bake it for less time in the oven.
You shared your mint chocolate chip cookies at the office and everyone LOVED them and begged you for your secret recipe. Since everyone had the same Betty Crocker book you just wrote that you followed page 152 and added mint chocolate chips and baked it for 10 minutes less. You put it on a bulletin board so everyone could see it but it didn’t have your name on it.
Next week Ethel brought in pecan cookies and put her modifications to page 152 on the board as well. Soon the whole board was filled with changes to page 152.
The manager came in and said we should really give these recipes to Betty Crocker. So he asked Yvonne, the secretary to summarize the board and write a letter to Betty Crocker. Yvonne put the popular cards together and sent them to Betty Crocker who put those new recipies in the 1962 edition.
Except the recipe is actually a model of your unconscious online behaviors. It is way of deciding what your next post should be, the next video you should watch, or the next ad you’ll be shown. Here’s two reasons why this will change the Internet as you know it
First, millions of people are contributing to the next version of the Betty Crocker cookbook. But the only person who gets paid is Betty Crocker. Your contributions are mandatory and your compensation is zero.
Betty Crocker is getting filthy rich over your contributions, but technically since your name was not on the recipe card so you will never see a dime. Your contributions are the legal property of Betty Crocker.
Betty Crocker might say that this is about protecting your privacy, in reality the more layers that anonymize your contribution the less claim you have to Betty Crocker’s profits. The less rights you have to correct misinformation about your online identity, the less rights you have information about you removed.
Legally, it’s not your information so Betty Crocker can do what she wants with it. Which leads me to my next point.
Data Rights
The second point is that there are different kinds of companies or cohorts, some are mom and pop shops in a poor black neighbourhood, others are the Ritz Carleton. This is not new, what you get shown online already depends on your race, gender, and income.
Using Ai to predict your socio-economic status is not new. In 2013 Facebook was granted a patent for inferring household income for users of a social networking system based on the sites that you visit and the place where you live.
Also, automatically opting you in to sending your data to big tech is not new as well.
For example, in 2018 Google starting automatically opting you into Chrome Browser logging when you’re logged into Google products such as Gmail.
In 2013 Gizmodo’s Kashmir Hill wrote about how Shadow profiles of people who had never created a Facebook account were discovered when a user downloaded their Facebook file, it included not just their friends’ visible contact information, but also their friends’ shadow contact information. The extent of the connections Facebook builds around its users is supposed to be visible only to the company itself, but in this case a bug revealed the truth: you’re information is being stored on Facebook even if you don’t have a Facebook account.
In the article you heard horrifying stories such as
- A man who years ago donated sperm to a couple, secretly, so they could have a child—only to have Facebook recommend the child as a person he should know. He still knows the couple but is not friends with them on Facebook.
- A social worker whose client called her by her nickname on their second visit, because she’d shown up in his People You May Know, despite their not having exchanged contact information.
- A woman whose father left her family when she was six years old—and saw his then-mistress suggested to her as a Facebook friend 40 years later.
- An attorney who wrote: “I deleted Facebook after it recommended as PYMK a man who was defense counsel on one of my cases. We had only communicated through my work email, which is not connected to my Facebook, which convinced me Facebook was scanning my work email.”
Facebook asks business owners to install a Facebook Pixel cookie tracker on many popular sites you visit so that your actions can be logged. Many modern browsers such as Brave block both trackers and ads. Since blogs make a lot of their revenue from these ads they require that unblock ads in order to watch articles or news.
https://gizmodo.com/how-facebook-figures-out-everyone-youve-ever-met-1819822691
What IS new is that you no longer know what cohort you are in and how many people are in your cohort. Why does this matter? If the cohort is small, say 10 or 100 people then it becomes a fingerprint that eventually leads back to you. If the cohort is big then it becomes a way of grouping people by broad categories such as gender, race, and income.
In the 2018 US election, we saw how mis-information targeted to poor black communities had an impact in growing voter apathy. The mostly non-black media didn’t notice these advertisements because they were served a very different set of advertisements on sites such as Facebook.
If the argument is made that big tech doesn’t actually know what cohort you are placed in they also avoid any liability for making suggestions that are potentially harmful to you.
So let’s say Betty Crocker releases a recipe that causes a medical reaction in a nut-free cookie company. Betty Crocker might argue that she is not liable because she never came up with the recipe to begin with, she’s only shared a recipe that an anonymous user came up with. When asked who that user was Betty Crocker could say that there’s no way to know who submitted it since they were all anonymous to begin with.
The anonymity gives protection to these companies. It’s going to be really hard to know what posts, videos, and ads are shown. It’s also going to be really hard to know where the suggestions come from and who they went to.
This matters to you because you won’t have the option to turn this off. You will not have the right to correct, delete, or change this data because it’s technically and legally not yours. Your data is now just anonymous modifications to a recipe.
Unconscious Advertising
Digital marketers and media buyers have known for years that training an Ai on a list of people who have bought from you in the past is far more effective than boosting a post based on characteristics from your profile such as age and gender alone.
The reason is that people don’t want their personal information on social media. They cleverly leave many of the personal identity fields blank believing that they have done their part to protect their privacy. However, unconscious input like milliseconds of watch time reveal the truth.
For example, even if Ernie didn’t write down his hobbies into his Facebook profile his watch time matches those of others that love fishing it’s pretty clear that fishing is Ernie’s hobby. So Ernie gets added to the fishing cohort even if he’s never liked, commented, or shared any fishing posts or joined any fishing groups.
This has been going on for years, so how is FLoC different that what has happened before?
The future of digital marketing and advertising is unconscious. Rather than boosting particular posts it will be about identifying a few of your ideal prospects on the platform and then getting an Ai to find people with similar online behaviors.
The input to the system will no longer be explicit personal characteristics like, age, gender, and hobbies.
When I first put in my birthday into Facebook there was an option to NOT include my year of birth. I thought it was great that there was a way to protect this information on Facebook. I was shocked years later when I saw that my birth year was on my profile even though I had never put it in myself. At the time I thought that Facebook may have got it from another site but I’ve since learned that’s not necessary. Ai can determine my age based on my friends and my online behavior.
Advertisers will be better at discriminatory ad targeting by creating cohorts of people with those characteristics and asking Ai to create look-alike audiences. For example, if I want people of a certain income level just put all of the people you know with that income level into a look-alike audience and then ask the Ai to find others that have the same browsing behavior. Even if people don’t put anything in their profile it should be able to find people that match that profile.
Ads in a world of FLoC will be more targeted, more discriminatory, and most importantly more effective. It will be like they are speaking specifically to you or people in your exact situation.
So what can you do about this?
Ai Parenting Insider
This week I shared with Insiders exclusive information that expands on Ethical Ai including reactions from my computer science peers. Leave a comment if you found today’s session helpful. If you apply for Ai Parenting Insider this week I’ll provide you with a code for a free screen time chart that makes it easier to see the quality of the screen time your family is spending today.
Thanks and I’ll see you next week where we will talk about skills your children can develop that won’t be replaced by Ai.
References
- David Temkin quote on Privacy https://blog.google/products/ads-commerce/a-more-privacy-first-web/
- FLoC in Chrome https://www.chromestatus.com/feature/5710139774468096
- Google FLoC is a Terrible Idea, EFF https://www.eff.org/deeplinks/2021/03/googles-floc-terrible-idea
- AI predicts Socio-Economic Status http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=8,600,797.PN.&OS=PN/8,600,797&RS=PN/8,600,797
- Chrome Logging since 2018 https://blog.cryptographyengineering.com/2018/09/23/why-im-leaving-chrome/
- Facebook Profile Download https://www.facebook.com/help/1701730696756992
- Facebook Friend Suggestions https://gizmodo.com/how-facebook-figures-out-everyone-youve-ever-met-1819822691