A good read from Tim O’Reilly. Could easily spend a day digging deeper into his references and further links. Some highlights for me (emphasis mine):
The essence of algorithm design is not to eliminate all error, but to make results robust in the face of error.
Accepting that mistakes are going to be made and that issues will arise. Robust is a great word for talking about design. Should be used more.
But the answer is not for Facebook to put journalists to work weeding out the good from the bad. It is to understand, in the same way that they’ve so successfully divined the features that lead to higher engagement, how to build algorithms that take into account “truth” as well as popularity.
This whole fake news issue has so many interesting angles and causes. Truth vs. Entertainment. Important vs. Popular. Making these decisions yourself, given time to consider and discuss is hard enough, let along writing code that will do the job for us. There are parallels here for me to the dangers of AI that feature at the end of Part 2 of the Wait But Why post about AI. It’s so hard to not anthropomorphise artificial intelligence, and expect it to have morals or understand nuances of what we intend. FB engineers must fully have expected and intend that popular posts would have been good. Crikey, imagine if the Facebook news feed algorithm were to develop artificial super intelligence – search this page for the story of Turry and see what I mean.
With Panda, Google took a big enough revenue hit via some partners that Google actually needed to disclose Panda as a material impact on an earnings call. But I believe it was the right decision to launch Panda, both for the long-term trust of our users and for a better ecosystem for publishers.
Here’s the rub. The bottom line. Whatever you call it. For Google to do ‘the right’ thing, to build trust, and make things better in the long-term it meant they would take a financial hit. They had to tell investors, warm them, practically seek permission to do the right thing. For all that’s said about right and wrong and true and popular, it’s misdirected to blame Facebook or even Breitbart when everyone is ultimately just trying to keep profits and projections pointing upwards. Any solution to fake news will only cost those who make it, and those who benefit from the trickle down economics. Remember, any of us with toy portfolios containing Google or Facebook, or a pension fund with equity investments in similar, then we’re in some way on the receiving end of click bait revenue.
Media in the Age of Algorithms – What’s The Future?