Book excerpt: The rise and fall of wisdom of the crowds

(This is an excerpt from drafts of my book, "Algorithms and Misinformation: Why Wisdom of the Crowds Failed the Internet and How to Fix It")

Wisdom of the crowds is the epiphany that combining many people's opinions is useful, often more useful than expert opinions.

Customer reviews summarize what thousands of people think of movies, books, and everything else you might want to buy. Customer reviews can be really useful for knowing if you want to buy something you've never tried before.

When you search the internet, what thousands of people clicked on before you helps determine what you see. Most of the websites on the internet are useless or scams; wisdom of the crowd filters all that out and helps you find what you need.

When you read the news online, you see news first that other people think is interesting. What people click determines what information you see about what's going on in the world.

Algorithms on the internet take the wisdom of the crowds to a gargantuan scale. Algorithms process all the data, summarizing it all down, until you get millions of people helping millions of people find what they need.

It sounds great, right? And it is. But once you use wisdom of the crowds, scammers come in. They see dollar signs in fooling those algorithms. Scammers profit from faking crowds.

When manipulated, wisdom of the crowds can promote scams, misinformation, and propaganda. Spammers clog up search engines until we can't see anything but scams. Online retailers are filled with bogus positive customer reviews of counterfeit and fraudulent items. The bad guys astroturf everything using fake crowds. Foreign operatives are able to flood the zone on social media with propaganda using thousands of fake accounts.

What we need is an internet that works for us. We need an internet that is useful and helpful, where we can find what we need without distractions and scams. Wisdom of the crowds and the algorithms that use wisdom of the crowds are the key to getting us there. But wisdom of the crowds can fail.

It's tricky to get right. Good intentions can lead to destructive outcomes. When executives tell their teams to optimize for clicks, they discover far too late that going down that path optimizes for scams and hate. When teams use big data, they're trying to make their algorithms work better, but they often end up sweeping up manipulated data that skews their results toward crap. Understanding why wisdom of the crowds fails and how to fix it is the key to getting us the internet we want.

The internet has come a long way. In the mid-1990s, it was just a few computer geeks. Nowadays, everyone in the world is online. There have been hard lessons learned along the way. These are the stories of unintended consequences.

Good intentioned efforts to tell teams to increase engagement caused misinformation and spam. Experimentation and A/B testing helped some teams help customers, but also accidentally sent some teams down dark paths of harming customers. Attempts to improve algorthms easily can go terribly wrong.

The internet has grown massively. During all of that growth, many internet companies struggled with figuring out how to make a real business. At first, Google had no revenue and no idea how to make money off web search. At first, Amazon had no profits and it was unclear if it ever would.

Almost always, people at tech companies had good intentions. We were scrambling to build the right thing. What we ended up building was not always the right thing. The surprising reason for this failure is what gets built depends not so much on the technology but the incentives people have.