Recently the company Facebook/Meta discussed exploring AI for its advertising. In part, this move may have been part of their plan to coexist alongside the “metaverse” — a term Zuckerberg did not invent but was happy to appropriate as a distinctly Facebook thing.
Apple announced back in 2021 with iOS update 14.5 that it would require all apps in its app store to display how the app would track a user’s data, and give the user the ability to opt out. Facebook lashed out with lawsuits, claiming Apple was discriminating against them due to it being known that the bulk of their revenue was tied to tracking users.
As the YouTube channel “New Money” covered succinctly, Meta would have to explore other avenues to continue feeding its advertising revenue model that’d served as the bulk of the company’s lifeblood.
(See New Money’s video below.)
Initially, less than 25% of users actually opted-in to targeted ads with these prompts. Facebook/Meta obviously felt that blow, since as a consequence the ads for all those users would be less relevant and less likely to connect.
Although it should be pretty telling that, as soon as there were prompts making it clear the ways certain apps wanted to track users, so few wanted to participate. People seemed happy to use Facebook for many years, knowing their data was being tracked in some ways.
Perhaps we all accepted it prior to this change because many of us weren’t sure exactly how the data was being tracked, or to exactly what degree, yet having it suddenly called out on-screen made it seem more objectionable.
Maybe knowing everyone’s data was also being logged made them feel better. “I’m no worse off than anyone else,” they may have thought.
Or perhaps they felt like they had no choice previously. All their friends and family were on Facebook so not being there didn’t seem viable, and so they would have to comply with whatever the data policies where.
Until there was a choice. Some choice anyway.
If the latter is true, it’s an interesting look at just how deeply Facebook had manipulated the public through social pressure to participate for over a decade.
That’s started to change, though. The Washington Post reported that in the last three months of 2021, Facebook lost a half million users.
It was the first quarter in Facebook’s nearly two decades run where such a decline had ever happened, by a wide margin.
Google Joined In With Their Own Data Transparency Moves.
Many people I spoke to saw Google’s sudden “me too” move to enable similar data transparency and opt-out features for Android devices as simply capitalizing on a popular stance. I can’t say it didn’t make sense to me, either.
If Apple was looking like a hero for championing user data privacy, Google would be damned if they didn’t catch some of that limelight, too.
It’s impossible to know the real motives of these companies, of course, but the timing was interesting. Google had taken its own flak for the last few years over being data-hungry, after all, and you might say that a move like this to regain some public goodwill was a smart one.
Whether only to craft a heroic image or not, Google throwing in with Apple on this stance was significant because there are over 3 billion active smartphone users around the world.
With both major smartphone platforms (Android and iPhone) making moves to allow users to better control their data, where does that leave a company whose product’s main focus is to harvest information about users for better advertising?
As New Money aptly pointed out in their video, these moves made it clearer than ever that controlling the hardware with which people access the internet is where the real power is.
The Augmented Reality/Virtual Reality Hype Train.
As you might expect at this point, Facebook and other companies saw tremendous value in creating hardware to leverage this new chapter of the internet.
Notably, Facebook teamed up with Ray-Ban to create smart sunglasses that reportedly can take calls, record video, play podcasts, and more. The glasses lack AR functionality, but you can bet it’s something Facebook will be adding soon.
It’ll be interesting to see how those glasses play out given the failure of “Google Glass” shortly after their public release in 2014.
A lot of people at the time saw Google Glass as creepy, citing the fact that since a user could take pictures and record video it’d allow a seemingly innocuous person to sit at a coffee shop, for instance, and record other people and their conversations.
An argument could be made that similar glasses might fare better now than in 2014, since our privacy has been invaded for so long now that an apathetic attitude of “what can you do?” seems more common.
I’ve certainly heard things like that huffed at me many a time. “Why bother? There’s nothing you can do to keep them out.”
On the other hand, perhaps not.
In 2021 when the news first broke of Facebook developing these smart glasses, various news sites cited the numerous privacy concerns. Andrew Bosworth of Facebook’s “Reality Labs” does not feel these glasses are in the same category as the Google Glass that came before them. He states, “This product has not been tried before because we’ve never had a design like this before.” (NY Times)
Evidently, he feels that these smart glasses being more focused on fashion than the tech within the glasses avoids the missteps of Google Glass.
What Would Meta’s AI Advertising Mean?
This actually wouldn’t be the first example of AI advertising. In 2018, Lexus notably released an ad that was completely written by an AI — one that had been given the directive to analyze 15 years of award-winning ads and analyze commonalities of what connected with audiences.
Facebook itself has been using some form of AI in its advertising for years.
What is most significant in this case is how much data can be analyzed in a short period of time, how many adjustments can be made, and how ads can be tailored in real time to maximize the likelihood that each ad connects.
Facebook’s “metaverse” is a perfect setup for AI ads. Since the entire digital space is controlled by Meta, theoretically every interaction a user makes within the metaverse can be logged and analyzed, and relevant ad opportunities could be theoretically served up to that user right within the metaverse in an unbroken way.
Imagine walking around some VR streets, looking at VR walls and seeing ad posters or billboards in your scenery. In all the ways the idea of being in an all new digital world seems cool, it’s exciting to advertisers for the same reasons.
Even in 2018 tech writers were talking about how Facebook’s AI could detect if a user were considering switching brands — even if that switch were in the future.
Humorously, as reported by Sam Biddle of The Intercept, a law professor described the advantages the AI provides as a machine learning protection racket. The algorithm can detect if a user is about to switch brands to a competitor, and then allow the brand paying the ad fees to quickly swoop in and turn them back.
Essentially, pay to play defensively and lock out the competition.
These machine learning algorithms are training themselves every day to become more effective at their given objectives. That includes the facial recognition software, Biddle writes.
FBLearner, Facebook’s primary learning algorithm for ads, essentially is fed a bunch of data about users, creates branched trees of its predicted outcomes, connects those assertions to similar users that are likely to follow similar behavior, and then delivers that ultimate report to corporations.
Those corporations can then tailor the ads to either reinforce the predicted behavior or try to prevent it.
But with anything that is that effective at predicting behavior, it’s understandable to see a growing concern about the algorithm’s ability to predict or even be used to influence other things, such as elections or the reactions to major events.