Capitol Hill tech hearings must focus on algorithms
US senators should probe on how the platforms manipulate users
The role of the world’s largest technology platforms in the spread of political propaganda, extremism and disinformation is extremely complex. That fact will be on display once again on Wednesday, as executives from Facebook, Twitter and possibly Google undergo another grilling by the US Senate intelligence committee over Russian meddling in the 2016 presidential election.
American politicians have so far struggled to ask the sort of pointed questions that would shed light on the role. The committee’s quizzing of Facebook CEO Mark Zuckerberg several months ago felt very much like a four-hour tech support call.
Meanwhile, Donald Trump has, in typical form, muddied the political waters by accusing Google of burying conservative news.
The president is wrong. While Silicon Valley leans left politically, Google and other platform technology companies do not employ evil geniuses to suppress Republican voices. Nor is it likely that executives from the top platforms were personally complicit in election manipulation.
But that does not mean that Google, Twitter, Facebook and other such companies do not have algorithmic biases. Indeed, their algorithms — which are designed to keep people on their sites for as long as possible — should be probed by senators on Wednesday.
While the word “algorithm” is now part of our common lexicon, few of us really understand what they are in the context of the websites we use every day. Algorithms are mathematical formulas that control how a site works — what search results come up, or what pieces of content are recommended to viewers. While users are online, their activities and preferences can be tracked and then monetised via targeted advertisements. The more time spent on Google or Facebook, the more the firms know about you. That detailed knowledge can then be sold to companies, individuals or even state actors. It is a highly effective business model, which is why those two companies alone now capture 85 per cent of digital advertising growth.
It is also a reason why the platform technology companies have been reluctant to open up the black box of their algorithms to allow the public and politicians to see how they work. Transparency would illuminate the disproportionate economic power they hold. It might also allow the public to see how the spread of unsavory content is part of the platform’s business model.
Platform firms are not trying, in some purposely nefarious way, to drive you towards cat videos or racist propaganda. But their algorithms are pushing many in exactly that direction.
That should be the focus of senate questioning. Facebook, Google, Twitter and others have responded to the outrage provoked by the spread of “fake news” and propaganda on their platforms by increasing the number of humans that police their sites for problematic content. That is an unsatisfactory solution.
The scale of the internet makes it impossible for humans to police everything. Moreover, these are technology firms, not media organisations. Their strength is not in curating content; it is in figuring out how to program algorithmic systems. Senators should ask how and why these algorithms are designed to keep us online longer, and whether that model might be fixed.
The answers may be complicated or vague. But complexity should not be a vehicle for obfuscation. The time has come for Silicon Valley to open the black box and reveal how the algorithms work. Trust in democracy, and capitalism, depends on it.