Category Archives: Linux

Should We Reveal AI Bias In Content Ranking Algorithms?

Pressure is rising. Will Google and Facebook need to hire armies of reviewers to risk further embarrassment? Their free pass to publish anything and everything is under attack, even if it’s still technically legal for them to do so. Rupert Murdoch is no stranger to the problem. Remember, Myspace under News Corp had to hire tons of reviewers too.

News Corp has said it wants to “focus the public and Congress on the anticompetitive behavior of the digital duopoly, especially as it adversely affects the news and information businesses.”

The difference now is, Google and Facebook should have access to amazing “machine learning” technology. Right? In theory, Google and Facebook reviewers armed with machine learning should be more effective than basic newspaper editors. Here’s what I’m thinking: Instead of wasting your time with moonshot ideas like self-driving cars and virtual reality, how about doing something realistic like policing your content?

It’s a battle of quality vs. quantity. My authority vs. your authority. A user-generated content army vs. a newspaper institution army in the battle for hearts, minds, rank and relevance. Many of us are on both sides of that equation, to a greater or lesser degree. I think newspapers have too much undeserved authority in this duopoly’s algorithms, but at least they review what they publish.

If Rupert Murdoch is looking to strike a bargain, his greatest weakness is undoubtedly PageRank. Google can press a few buttons and all of Rupert’s newspapers drop back in authority to page 2 in the SERPs. Same for Facebook, which has an even more abstract ranking system we don’t even have a name for. Just from that perspective, the scale of this duopoly’s power is unnerving. Assuming the public has the ability to do so, should we nudge or “tweak” the tech titans’ algorithmic bias to be more fair? Or is the fairness of an algorithm subjective? Do we strictly optimize for more interactions and profits? After all, clicks on ads keep these companies in business.

To be fair, Google and Facebook have strict rules in terms of what can be advertised, and they decide in what context those ads can be viewed. The difference here is, the ad business is regulated by governments. The debate here is about how content should be ranked. And the new trend is, content (marketing) is the new advertising. So there’s even more at stake than before, in terms of dollars. At the same time, every special interest group on the planet is now concerned about “bias” in these ranking algorithms, as they come to impact more and more of our lives.

Google and Facebook will tell you they are very important, and that they have great influence over society in terms of how people think and act, and this is due in part to how they rank and rate content. Given that, do these same tech CEOs look inward for moral authority? Do they trust their own flavors of wisdom to guide such powerful algorithms in the right direction? Do they even want to live in this world they helped create?

Can they sleep at night, when a few lines of code in an algorithm have the power to affect the lives of billions of people? I don’t really blame superpower countries, or ordinary people, for wanting more control over their own destiny. How are these algorithms used, for or against us? What’s the score? Where do I stand? What I do know, people and countries trust open source software like Linux, and simultaneously distrust Microsoft Windows and other “black box” hardware and software.

That’s how free speech gets entangled with algorithmic bias. Curation is content too. Maybe I don’t want your version of reality. Google and Facebook are compelling versions of reality. You can argue Facebook and Google algorithms can’t completely suppress information, and that is true. But there’s no arguing they do have unprecedented power to magnify and silence the voice of whomever they choose.

Don’t get me wrong, I’m not in favor of open sourcing everything. But with far-reaching technology like Google and Facebook, datacenters upon datacenters, with deep tentacles thirstily feasting on every aspect of billions of lives, recording our speech, facial expressions, every motion captured by gyroscope–where do we draw the line?

Google surprised the audience with a new camera called Google Clips. It automatically takes photos as you go throughout your day. An artificial intelligence engine inside looks for smiles, people you care about, dogs chasing their tails, and automatically captures the moment. Google says it will get smarter over time.

Yes, please smile or I lose points! In terms of the push for artificial intelligence, the data wants to run wild. Such a machine behind it, Google’s untouchable black box, aka Fred, will naturally want more and more control over what it considers important and authoritative, as that is core to ranking people, institutions, content, countries, religions, etc. You better be smiling! Let’s not forget Bing either, which now claims 33 percent search market share in the US.