3 ms·
Let me tell you a story. There once was a company that made a social media platform for sharing video content. The platform was monetized by ads. The company
by CivBase 6y ago
Let me tell you a story.
There once was a company that made a social media platform for sharing video content. The platform was monetized by ads. The company noticed that its users watched more ads so long as they were being shown content which appealed to their personal preferences and opinions. The company dubbed this phenomena "engagement" and quickly introduced an algorithm to maximize engagement by constantly suggesting content which appealed to each individual user.
People loved the company's social media platform and it quickly grew larger than any of its alternatives. Luckily, the company already controlled the world's most popular search engine, web browser, and mobile operating system, so it was easy to make their platform and its addicting algorithm as accessible as possible. A few competitors still existed, but none came anywhere close to the scale and accessibility of the company's platform.
But something happened.
The algorithm which helped the platform grow and made so much money was having an adverse affect on the platform's users. The users were only being exposed to content they liked, which helped establish echo chambers. People were coddled by content which helped reinforce their existing opinions - never challenge them. As people became more and more established in their echo chambers, misinformation began to spread like wild fire. Users were desperate and the algorithm rewarded creators who provided content to help them avoid challenging their own opinions - regardless of whether it was true or not.
What was the company to do? The algorithm was too profitable for them to simply turn it off - especially when it came to political content. They tried demonetizing content about particular subjects and adding "fact check" bubbles to combat the spread of misinformation but it wasn't enough. They eventually started trying to identify and remove the content with misinformation altogether, but even that wasn't enough.
But then something amazing happens and th- oh my look at the time! I need to go now. Maybe we can finish this story some other day.
- esja 6y agoOne day it will be commonly accepted that these ML recommenders have played the same role in this social crisis as the Gaussian copula did in the financial crisis.