Why is Wikipedia mostly OK, but why do so many comments at the end of newspaper articles make you weep for the future of humanity – when both are open for everyone to write?

Tay_bot_logo

Earlier this week I got to listen to Jimmy Wales of Wikipedia at Khazanah Megatrends Forum. He reflected on, among other things, the openness of the system of knowledge production in Wikipedia, and its sustainability. Just the day earlier I had gotten to speak to yet others about an early Microsoft experiment on openness and Artificial Intelligence: Tay the Internet Chatbot. Naturally, I think these story arcs reflect also on the larger social science question of the workings of open and closed societies, at least as traditionally conceptualised.

 

2017.10.02-MOKL-Khazanah-Megatrends-Forum-2017-Jimmy.Wales

Many find appealing the hypothesis that open, individual-oriented, bottom-up societies achieve outcomes that, even if not point-wise optimal, are at least dynamically robust and resilient. Open systems might suffer from bad shocks – financial crisis, Trump – but they recover well and get back on track quickly. In contrast, according to this reasoning, closed autocracies – even if they achieve Disneyland-like success and performance in the short to medium term – are fragile, and eventually fail. Without moving to greater openness, those parts of East Asia that have seen economic success should prepare themselves for the risk of a collapse of Mugabe-esque proportion.

In this view, across the universe of closed systems, one might see diversity in outcome – both failure and success – in the short to medium term. But open systems should systematically display success in the long term, because they can flexibly adapt to shocks.

Openness and individual empowerment are qualities in themselves desirable and deserving of aspiration. The question here is not about those intrinsic attractions; it is instead about their implications for systemic outcomes. What are the facts: do all open systems show robustness and resilience?

Why is Wikipedia, in the main, sensible, but comments at the end of even mainstream newspaper articles make you weep for the future of humanity? Both systems of knowledge production are, to an untutored eye, equally open. Everybody can edit Wikipedia; newspapers that try to restrict the worst of excesses from their readers’ comments either slow to a crawl or end up, in other ways, still unable to stem the flood of vitriol. How did Internet trolls end up keying in so much venom-driven input to drive Tay the Microsoft AI chatbot to become a being of hate and virulent racism and sexism?

In the first, the user community is empowered as a self-organised but emergent and recognisable entity; in the latter two, not so, but instead remained a ragtag group of referendum-ish devolved units. What rules made for one outcome, what rules got us to the other? Levels of hierarchy are evidently not inconsistent with open systems, despite the appeal of naive one-person, one-vote referendum.

Saying just “rules-based open system” doesn’t seem to me a particularly sharp tool as a separating surface for what, on the one side, works for social systems and what, on the other, fails. It seems to me that social science – politics, International Relations, economics – needs a Grand Unifying Theorem, something larger even than the Fundamental Theorems of Welfare Economics, against which we can set and study real-world counter-examples.