Say the words “Cambridge Analytica” to a western politician — or anyone clutching a smartphone — and they will probably wince. After all, this year some startling revelations have tumbled out about how that British consulting company harvested social media and consumer data to build manipulative political and business campaigns.
This week Britain’s Information Commissioner’s Office imposed a record £500,000 fine on Facebook for letting Cambridge Analytica use an app to access up to 87m user profiles without permission. The ICO intends to bring criminal proceedings against Cambridge Analytica’s parent over its failure to deal properly with the regulator’s enforcement notice.
It makes for lurid headlines: Cambridge Analytica did work for Donald Trump’s 2016 presidential campaign. Western intelligence groups say that Russia used social media to meddle in that election and the UK’s Brexit referendum.
If you want to take a wider view of the underlying policy issues, ponder the parallels between this year’s data harvesting tale and the growth in the 2000s of financial innovations such as credit derivatives. The parallel might initially sound odd. Credit derivatives allow financiers to make bets about whether people, companies or countries will default on their debt. Data scientists at Cambridge Analytica amass consumer data to study (and manipulate) humans. Those tasks do not seem similar.
But the parallel — and policy lesson — lies in how politicians handle innovation; or mishandle it. The idea of credit derivatives first cropped up in the late 20th century when some young(ish) financiers at places such as JPMorgan embarked on frenetic innovation. Most non-bankers had no idea what the whizz-kids were doing, or how their inventions might eventually change corporate and mortgage debt. For their part, the finance geeks were not hiding their innovations or being deliberately malicious; they told themselves (and others) that their inventions would improve the financial system, while also making them rich.
But credit derivatives and other financial innovations were widely ignored because they seemed so “nerdy” and mind-numbingly dull. Politicians and voters had no incentive to ask hard questions because they were enjoying the cheap mortgages and credit cards that the derivatives helped make possible. Regulators were largely toothless because the whizz kids were creating financial instruments that straddled national borders, regulatory silos and outdated laws. The power rested with the geeks until their innovations were abused in a way that contributed to a financial crisis.
Fast forward to now. Did this saga involve a team of young(ish) geeks who felt intoxicated by the intellectual thrill of fast-paced innovation? Yes, although their new intellectual frontier was in data science, not finance. Did those geeks want to get rich? Definitely. As for regulation, the innovation in computer science, like that in finance, has jumped rapidly across national borders and around existing laws.
Politicians and voters also failed to demand proper oversight. The world of big data seems as geeky and dull to outsiders now as derivatives did back then. Consumers like “free” internet services as much as they enjoyed cheap mortgages. No one wanted to think about the hidden costs.
Maybe this will now change. The Facebook and Cambridge Analytica scandal has finally forced politicians to debate the risks of big data — just as the 2008 financial crisis sparked public scrutiny of finance. Regulators are belatedly crafting policy responses. Consumers are becoming (a little) more savvy about the cost of “free”.
It is possible to hope that this political hand-wringing will eventually create a healthier world of data science, just as the 2008 financial crisis brought the banking system a (little more) under control. Innovation is a double-edged sword. Big data can be misused by unscrupulous politicians and businesses, just as derivatives can be dangerous in the hands of greedy or reckless financiers (mixing them with subprime mortgages was a very bad idea). But these abuses do not make the core idea of data science — or derivatives — “bad”; on the contrary, these advances can sometimes be very useful tools.
The key lesson is that innovation only delivers real benefits when there is a proper system of ethics, political oversight and up-to-date laws.
Politicians and consumers must scrutinise the geeks in a timely way. Particularly when those geeks come bearing consumer gifts that seem temptingly cheap, easy to access or “free”.