Is Facebook's "10 Year Challenge" A Ploy To Teach Facial Recognition Algorithms

If you have been on social media recently, you’ve likely been subjected to the "10 year challenge", a growing trend of people posting photos of themselves from 10 years ago and comparing themselves to current photos. Like most social media "challenges", it isn’t much of a "challenge" at all, but rather an excuse for insecure social media users to post more Snapchat and Instagram-filtered photos of themselves in hopes of proving to the world that their life has deep meaning beyond, well, being constantly on Facebook, SnapChat or Instagram. 

But one author and expert in the field had a different take: she raised an interesting conspiratorial view that stemmed from a semi sarcastic tweet she put out on January 12. Speaker, entrepreneur and author Kate O’Neill wrote the following:

She then took the time to write an article for Wired that "goes down the rabbit hole" a bit and runs with the conspiracy theory.

O'Neill's has a background in integrated experience strategy and human-centric digital transformation as a result of her "more than 20 years of experience and entrepreneurship leading innovations across technology, marketing, and operations, developing human-centric, data-guided, and brand-aligned growth and retention strategies for companies of all sizes, from startups to Fortune 500s."

She claims that if you wanted to train a facial recognition algorithm on age related characteristics and age progression, you would want a lot of people's photos and you would want to know that they were all taken a fixed number of years apart – 10 years, in this case.

There’s certainly the counter-argument that you could mine Facebook data as it stands, making the "challenge" unnecessary, but photos are sometimes put up out of order and often feature images of items that are much more than users: word images, cartoons, patterns, and others. The EXIF data on these photos is also unreliable, as people have uploaded and scanned photographs from different eras at different times.

And so it would help if there was a clean, simple and rigorously labeled set of "then and now" photographs, much like we are seeing with the "10 year challenge". 

O'Neill makes a cogent and objective point: thanks to this meme challenge, there’s an extremely large data set of carefully picked photos of people from roughly 10 years ago and now.

Some people argue that there is too much useless data for this challenge to be useful. But the article makes the argument that people (and arguably a company using this data for nefarious purposes) would know you're supposed to place more trust in the validity of data earlier on in a trend. This just means that someone would have to be more likely to mine the data that started coming out at the beginning of the challenge. And, by now, algorithms are smart enough to separate human faces from the joke memes that people are putting up, like photos of their cats and dogs.

Facebook denies having any involvement. They told Wired: "This is a user-generated meme that went viral on its own. Facebook did not start this trend, and the meme uses photos that already exist on Facebook. Facebook gains nothing from this meme (besides reminding us of the questionable fashion trends of 2009). As a reminder, Facebook users can choose to turn facial recognition on or off at any time.”

Of course, the idea that the information is being used for nefarious purposes hasn't been proven and is certainly still a conspiracy theory. But a key point O'Neill tries to get across in her piece certainly isn't: "humans are the richest data sources for most of the technology emerging in the world". 

You can read the full piece here